feat: addressed comments

frontend/src/constants/localStorage.ts

@@ -37,4 +37,5 @@ export enum LOCALSTORAGE {
 	SHOW_FREQUENCY_CHART = 'SHOW_FREQUENCY_CHART',
 	DISSMISSED_COST_METER_INFO = 'DISMISSED_COST_METER_INFO',
 	DISMISSED_API_KEYS_DEPRECATION_BANNER = 'DISMISSED_API_KEYS_DEPRECATION_BANNER',
+	LICENSE_KEY_CALLOUT_DISMISSED = 'LICENSE_KEY_CALLOUT_DISMISSED',
 }

frontend/src/container/GeneralSettings/GeneralSettings.tsx

@@ -38,6 +38,7 @@ import {
 } from 'types/api/settings/getRetention';
 import { USER_ROLES } from 'types/roles';
 
+import LicenseRowDismissibleCallout from './LicenseKeyRow/LicenseRowDismissibleCallout/LicenseRowDismissibleCallout';
 import Retention from './Retention';
 import StatusMessage from './StatusMessage';
 import { ActionItemsContainer, ErrorText, ErrorTextContainer } from './styles';
@@ -683,7 +684,12 @@ function GeneralSettings({
 					{showCustomDomainSettings && activeLicense?.key && (
 						<div className="custom-domain-card-divider" />
 					)}
-					{activeLicense?.key && <LicenseKeyRow />}
+					{activeLicense?.key && (
+						<>
+							<LicenseKeyRow />
+							<LicenseRowDismissibleCallout />
+						</>
+					)}
 				</div>
 			)}
 

frontend/src/container/GeneralSettings/LicenseKeyRow/LicenseRowDismissibleCallout/LicenseRowDismissible.styles.scss

@@ -0,0 +1,31 @@
+.license-key-callout {
+	margin: var(--spacing-4) var(--spacing-6);
+	width: auto;
+
+	.license-key-callout__description {
+		display: flex;
+		align-items: baseline;
+		gap: var(--spacing-2);
+		min-width: 0;
+		flex-wrap: wrap;
+		font-size: 13px;
+	}
+
+	.license-key-callout__link {
+		display: inline-flex;
+		align-items: center;
+		padding: 1px 5px;
+		border-radius: 2px;
+		background: var(--callout-primary-background);
+		color: var(--callout-primary-description);
+		font-family: 'SF Mono', 'Fira Code', 'Fira Mono', monospace;
+		font-size: var(--paragraph-base-400-font-size);
+		text-decoration: none;
+
+		&:hover {
+			background: var(--callout-primary-border);
+			color: var(--callout-primary-icon);
+			text-decoration: none;
+		}
+	}
+}

frontend/src/container/GeneralSettings/LicenseKeyRow/LicenseRowDismissibleCallout/LicenseRowDismissibleCallout.tsx

@@ -0,0 +1,87 @@
+import { useState } from 'react';
+import { Link } from 'react-router-dom';
+import { Callout } from '@signozhq/callout';
+import getLocalStorageApi from 'api/browser/localstorage/get';
+import setLocalStorageApi from 'api/browser/localstorage/set';
+import { FeatureKeys } from 'constants/features';
+import { LOCALSTORAGE } from 'constants/localStorage';
+import ROUTES from 'constants/routes';
+import { useGetTenantLicense } from 'hooks/useGetTenantLicense';
+import { useAppContext } from 'providers/App/App';
+import { USER_ROLES } from 'types/roles';
+
+import './LicenseRowDismissible.styles.scss';
+
+function LicenseRowDismissibleCallout(): JSX.Element | null {
+	const [isCalloutDismissed, setIsCalloutDismissed] = useState<boolean>(
+		() =>
+			getLocalStorageApi(LOCALSTORAGE.LICENSE_KEY_CALLOUT_DISMISSED) === 'true',
+	);
+
+	const { user, featureFlags } = useAppContext();
+	const { isCloudUser } = useGetTenantLicense();
+
+	const isAdmin = user.role === USER_ROLES.ADMIN;
+	const isEditor = user.role === USER_ROLES.EDITOR;
+
+	const isGatewayEnabled =
+		featureFlags?.find((feature) => feature.name === FeatureKeys.GATEWAY)
+			?.active || false;
+
+	// Service accounts are only accessible to admins
+	const hasServiceAccountsAccess = isAdmin;
+
+	// Ingestion settings are accessible to:
+	// - Cloud users when gateway is not enabled
+	// - Admin/Editor when gateway is enabled
+	const hasIngestionAccess =
+		(isCloudUser && !isGatewayEnabled) ||
+		(isGatewayEnabled && (isAdmin || isEditor));
+
+	const handleDismissCallout = (): void => {
+		setLocalStorageApi(LOCALSTORAGE.LICENSE_KEY_CALLOUT_DISMISSED, 'true');
+		setIsCalloutDismissed(true);
+	};
+
+	return !isCalloutDismissed ? (
+		<Callout
+			type="info"
+			size="small"
+			showIcon
+			dismissable
+			onClose={handleDismissCallout}
+			className="license-key-callout"
+			description={
+				<div className="license-key-callout__description">
+					This is <strong>NOT</strong> your ingestion or Service account key.
+					{(hasServiceAccountsAccess || hasIngestionAccess) && (
+						<>
+							{' '}
+							Find your{' '}
+							{hasServiceAccountsAccess && (
+								<Link
+									to={ROUTES.SERVICE_ACCOUNTS_SETTINGS}
+									className="license-key-callout__link"
+								>
+									Service account here
+								</Link>
+							)}
+							{hasServiceAccountsAccess && hasIngestionAccess && ' and '}
+							{hasIngestionAccess && (
+								<Link
+									to={ROUTES.INGESTION_SETTINGS}
+									className="license-key-callout__link"
+								>
+									Ingestion key here
+								</Link>
+							)}
+							.
+						</>
+					)}
+				</div>
+			}
+		/>
+	) : null;
+}
+
+export default LicenseRowDismissibleCallout;

frontend/src/container/GeneralSettings/LicenseKeyRow/LicenseRowDismissibleCallout/__tests__/LicenseRowDismissibleCallout.test.tsx

@@ -0,0 +1,219 @@
+import { FeatureKeys } from 'constants/features';
+import { LOCALSTORAGE } from 'constants/localStorage';
+import ROUTES from 'constants/routes';
+import { useGetTenantLicense } from 'hooks/useGetTenantLicense';
+import { render, screen, userEvent } from 'tests/test-utils';
+import { USER_ROLES } from 'types/roles';
+
+import LicenseRowDismissibleCallout from '../LicenseRowDismissibleCallout';
+
+const getDescription = (): HTMLElement =>
+	screen.getByText(
+		(_, el) =>
+			el?.classList?.contains('license-key-callout__description') ?? false,
+	);
+
+const queryDescription = (): HTMLElement | null =>
+	screen.queryByText(
+		(_, el) =>
+			el?.classList?.contains('license-key-callout__description') ?? false,
+	);
+
+jest.mock('hooks/useGetTenantLicense', () => ({
+	useGetTenantLicense: jest.fn(),
+}));
+
+const mockLicense = (isCloudUser: boolean): void => {
+	(useGetTenantLicense as jest.Mock).mockReturnValue({
+		isCloudUser,
+		isEnterpriseSelfHostedUser: !isCloudUser,
+		isCommunityUser: false,
+		isCommunityEnterpriseUser: false,
+	});
+};
+
+const renderCallout = (
+	role: string,
+	isCloudUser: boolean,
+	gatewayActive: boolean,
+): void => {
+	mockLicense(isCloudUser);
+	render(
+		<LicenseRowDismissibleCallout />,
+		{},
+		{
+			role,
+			appContextOverrides: {
+				featureFlags: [
+					{
+						name: FeatureKeys.GATEWAY,
+						active: gatewayActive,
+						usage: 0,
+						usage_limit: -1,
+						route: '',
+					},
+				],
+			},
+		},
+	);
+};
+
+describe('LicenseRowDismissibleCallout', () => {
+	beforeEach(() => {
+		localStorage.clear();
+		jest.clearAllMocks();
+	});
+
+	describe('callout content per access level', () => {
+		it.each([
+			{
+				scenario: 'viewer, non-cloud, gateway off — base text only, no links',
+				role: USER_ROLES.VIEWER,
+				isCloudUser: false,
+				gatewayActive: false,
+				serviceAccountLink: false,
+				ingestionLink: false,
+				expectedText: 'This is NOT your ingestion or Service account key.',
+			},
+			{
+				scenario: 'admin, non-cloud, gateway off — service accounts link only',
+				role: USER_ROLES.ADMIN,
+				isCloudUser: false,
+				gatewayActive: false,
+				serviceAccountLink: true,
+				ingestionLink: false,
+				expectedText:
+					'This is NOT your ingestion or Service account key. Find your Service account here.',
+			},
+			{
+				scenario: 'viewer, cloud, gateway off — ingestion link only',
+				role: USER_ROLES.VIEWER,
+				isCloudUser: true,
+				gatewayActive: false,
+				serviceAccountLink: false,
+				ingestionLink: true,
+				expectedText:
+					'This is NOT your ingestion or Service account key. Find your Ingestion key here.',
+			},
+			{
+				scenario: 'admin, cloud, gateway off — both links',
+				role: USER_ROLES.ADMIN,
+				isCloudUser: true,
+				gatewayActive: false,
+				serviceAccountLink: true,
+				ingestionLink: true,
+				expectedText:
+					'This is NOT your ingestion or Service account key. Find your Service account here and Ingestion key here.',
+			},
+			{
+				scenario: 'admin, non-cloud, gateway on — both links',
+				role: USER_ROLES.ADMIN,
+				isCloudUser: false,
+				gatewayActive: true,
+				serviceAccountLink: true,
+				ingestionLink: true,
+				expectedText:
+					'This is NOT your ingestion or Service account key. Find your Service account here and Ingestion key here.',
+			},
+			{
+				scenario: 'editor, non-cloud, gateway on — ingestion link only',
+				role: USER_ROLES.EDITOR,
+				isCloudUser: false,
+				gatewayActive: true,
+				serviceAccountLink: false,
+				ingestionLink: true,
+				expectedText:
+					'This is NOT your ingestion or Service account key. Find your Ingestion key here.',
+			},
+		])(
+			'$scenario',
+			({
+				role,
+				isCloudUser,
+				gatewayActive,
+				serviceAccountLink,
+				ingestionLink,
+				expectedText,
+			}) => {
+				renderCallout(role, isCloudUser, gatewayActive);
+
+				const description = getDescription();
+				expect(description).toBeInTheDocument();
+				expect(description).toHaveTextContent(expectedText);
+
+				if (serviceAccountLink) {
+					expect(
+						screen.getByRole('link', { name: /Service account here/ }),
+					).toBeInTheDocument();
+				} else {
+					expect(
+						screen.queryByRole('link', { name: /Service account here/ }),
+					).not.toBeInTheDocument();
+				}
+
+				if (ingestionLink) {
+					expect(
+						screen.getByRole('link', { name: /Ingestion key here/ }),
+					).toBeInTheDocument();
+				} else {
+					expect(
+						screen.queryByRole('link', { name: /Ingestion key here/ }),
+					).not.toBeInTheDocument();
+				}
+			},
+		);
+	});
+
+	describe('Link routing', () => {
+		it('should link to service accounts settings', () => {
+			renderCallout(USER_ROLES.ADMIN, false, false);
+
+			const link = screen.getByRole('link', {
+				name: /Service account here/,
+			}) as HTMLAnchorElement;
+
+			expect(link.getAttribute('href')).toBe(ROUTES.SERVICE_ACCOUNTS_SETTINGS);
+		});
+
+		it('should link to ingestion settings', () => {
+			renderCallout(USER_ROLES.VIEWER, true, false);
+
+			const link = screen.getByRole('link', {
+				name: /Ingestion key here/,
+			}) as HTMLAnchorElement;
+
+			expect(link.getAttribute('href')).toBe(ROUTES.INGESTION_SETTINGS);
+		});
+	});
+
+	describe('Dismissal functionality', () => {
+		it('should hide callout when dismiss button is clicked', async () => {
+			const user = userEvent.setup();
+			renderCallout(USER_ROLES.ADMIN, false, false);
+
+			expect(getDescription()).toBeInTheDocument();
+
+			await user.click(screen.getByRole('button'));
+
+			expect(queryDescription()).not.toBeInTheDocument();
+		});
+
+		it('should persist dismissal in localStorage', async () => {
+			const user = userEvent.setup();
+			renderCallout(USER_ROLES.ADMIN, false, false);
+
+			await user.click(screen.getByRole('button'));
+
+			expect(
+				localStorage.getItem(LOCALSTORAGE.LICENSE_KEY_CALLOUT_DISMISSED),
+			).toBe('true');
+		});
+
+		it('should not render when localStorage dismissal is set', () => {
+			localStorage.setItem(LOCALSTORAGE.LICENSE_KEY_CALLOUT_DISMISSED, 'true');
+			renderCallout(USER_ROLES.ADMIN, false, false);
+
+			expect(queryDescription()).not.toBeInTheDocument();
+		});
+	});
+});

pkg/querybuilder/collision.go

@@ -207,16 +207,23 @@ func AdjustKey(key *telemetrytypes.TelemetryFieldKey, keys map[string][]*telemet
 		indexes := []telemetrytypes.JSONDataTypeIndex{}
 		fieldContextsSeen := map[telemetrytypes.FieldContext]bool{}
 		dataTypesSeen := map[telemetrytypes.FieldDataType]bool{}
+		jsonTypesSeen := map[string]*telemetrytypes.JSONDataType{}
 		for _, matchingKey := range matchingKeys {
 			materialized = materialized && matchingKey.Materialized
 			fieldContextsSeen[matchingKey.FieldContext] = true
 			dataTypesSeen[matchingKey.FieldDataType] = true
+			if matchingKey.JSONDataType != nil {
+				jsonTypesSeen[matchingKey.JSONDataType.StringValue()] = matchingKey.JSONDataType
+			}
 			indexes = append(indexes, matchingKey.Indexes...)
 		}
 		for _, matchingKey := range contextPrefixedMatchingKeys {
 			materialized = materialized && matchingKey.Materialized
 			fieldContextsSeen[matchingKey.FieldContext] = true
 			dataTypesSeen[matchingKey.FieldDataType] = true
+			if matchingKey.JSONDataType != nil {
+				jsonTypesSeen[matchingKey.JSONDataType.StringValue()] = matchingKey.JSONDataType
+			}
 			indexes = append(indexes, matchingKey.Indexes...)
 		}
 		key.Materialized = materialized
@@ -241,6 +248,15 @@ func AdjustKey(key *telemetrytypes.TelemetryFieldKey, keys map[string][]*telemet
 				break
 			}
 		}
+
+		if len(jsonTypesSeen) == 1 && key.JSONDataType == nil {
+			// all matching keys have same JSON data type, use it
+			for _, jt := range jsonTypesSeen {
+				actions = append(actions, fmt.Sprintf("Adjusting key %s to have JSON data type %s", key, jt.StringValue()))
+				key.JSONDataType = jt
+				break
+			}
+		}
 	}
 
 	return actions

pkg/telemetrylogs/field_mapper.go

@@ -276,12 +276,11 @@ func (m *fieldMapper) FieldFor(ctx context.Context, tsStart, tsEnd uint64, key *
 					continue
 				}
 
-				if key.FieldDataType == telemetrytypes.FieldDataTypeUnspecified {
+				if key.JSONDataType == nil {
 					return "", qbtypes.ErrColumnNotFound
 				}
 
-				jdt := key.GetJSONDataType()
-				if key.KeyNameContainsArray() && !jdt.IsArray {
+				if key.KeyNameContainsArray() && !key.JSONDataType.IsArray {
 					return "", errors.NewInvalidInputf(errors.CodeInvalidInput, "FieldFor not supported for nested fields; only supported for flat paths (e.g. body.status.detail) and paths of Array type: %s(%s)", key.Name, key.FieldDataType)
 				}
 

pkg/telemetrylogs/json_condition_builder.go

@@ -173,9 +173,29 @@ func (c *jsonConditionBuilder) terminalIndexedCondition(node *telemetrytypes.JSO
 		return "", errors.NewInternalf(CodeArrayNavigationFailed, "can not build index condition for array field %s", fieldPath)
 	}
 
-	indexedExpr := assumeNotNull(fmt.Sprintf("dynamicElement(%s, '%s')", fieldPath, node.TerminalConfig.ElemType.StringValue()))
+	elemType := node.TerminalConfig.ElemType
+	dynamicExpr := fmt.Sprintf("dynamicElement(%s, '%s')", fieldPath, elemType.StringValue())
+	indexedExpr := assumeNotNull(dynamicExpr)
+
+	// switch the operator and value for exists and not exists
+	switch operator {
+	case qbtypes.FilterOperatorExists:
+		operator = qbtypes.FilterOperatorNotEqual
+		value = getEmptyValue(elemType)
+	case qbtypes.FilterOperatorNotExists:
+		operator = qbtypes.FilterOperatorEqual
+		value = getEmptyValue(elemType)
+	default:
+		// do nothing
+	}
+
 	indexedExpr, formattedValue := querybuilder.DataTypeCollisionHandledFieldName(node.TerminalConfig.Key, value, indexedExpr, operator)
-	return c.applyOperator(sb, indexedExpr, operator, formattedValue)
+	cond, err := c.applyOperator(sb, indexedExpr, operator, formattedValue)
+	if err != nil {
+		return "", err
+	}
+
+	return cond, nil
 }
 
 // buildPrimitiveTerminalCondition builds the condition if the terminal node is a primitive type
@@ -184,31 +204,32 @@ func (c *jsonConditionBuilder) buildPrimitiveTerminalCondition(node *telemetryty
 	fieldPath := node.FieldPath()
 	conditions := []string{}
 
-	// Utilize indexes when available, except for EXISTS/NOT EXISTS checks.
-	// Indexed columns always store a default empty value for absent fields (e.g. "" for strings,
-	// 0 for numbers), so using the index for existence checks would incorrectly exclude rows where
-	// the field genuinely holds the empty/zero value.
+	// utilize indexes for the condition if available
 	//
-	// Note: indexing is also skipped for Array Nested fields because they cannot be indexed.
+	// Note: Indexing code doesn't get executed for Array Nested fields because they can not be indexed
 	indexed := slices.ContainsFunc(node.TerminalConfig.Key.Indexes, func(index telemetrytypes.JSONDataTypeIndex) bool {
 		return index.Type == node.TerminalConfig.ElemType
 	})
-	isExistsCheck := operator == qbtypes.FilterOperatorExists || operator == qbtypes.FilterOperatorNotExists
-	if node.TerminalConfig.ElemType.IndexSupported && indexed && !isExistsCheck {
+	if node.TerminalConfig.ElemType.IndexSupported && indexed {
 		indexCond, err := c.terminalIndexedCondition(node, operator, value, sb)
 		if err != nil {
 			return "", err
 		}
-		// With a concrete non-zero value the index condition is self-contained.
+		// if qb has a definitive value, we can skip adding a condition to
+		// check the existence of the path in the json column
 		if value != nil && value != getEmptyValue(node.TerminalConfig.ElemType) {
 			return indexCond, nil
 		}
-		// The value is nil or the type's zero/empty value. Because indexed columns always store
-		// that zero value for absent fields, the index alone cannot distinguish "field is absent"
-		// from "field exists with zero value". Append a path-existence check (IS NOT NULL) as a
-		// second condition and AND them together.
+
 		conditions = append(conditions, indexCond)
-		operator = qbtypes.FilterOperatorExists
+
+		// Switch operator to EXISTS except when operator is NOT EXISTS since
+		// indexed paths on assumedNotNull, indexes will always have a default
+		// value so we flip the operator to Exists and filter the rows that
+		// actually have the value
+		if operator != qbtypes.FilterOperatorNotExists {
+			operator = qbtypes.FilterOperatorExists
+		}
 	}
 
 	var formattedValue = value
@@ -218,15 +239,20 @@ func (c *jsonConditionBuilder) buildPrimitiveTerminalCondition(node *telemetryty
 
 	fieldExpr := fmt.Sprintf("dynamicElement(%s, '%s')", fieldPath, node.TerminalConfig.ElemType.StringValue())
 
-	// For non-nested paths with a negative comparison operator (e.g. !=, NOT LIKE, NOT IN),
-	// wrap in assumeNotNull so ClickHouse treats absent paths as the zero value rather than NULL,
-	// which would otherwise cause them to be silently dropped from results.
-	// NOT EXISTS is excluded: we want a true NULL check there, not a zero-value stand-in.
+	// if operator is negative and has a value comparison i.e. excluding EXISTS and NOT EXISTS, we need to assume that the field exists everywhere
 	//
-	// Note: for nested array paths, buildAccessNodeBranches already inverts the operator before
-	// reaching here, so IsNonNestedPath() guards against double-applying the wrapping.
-	if node.IsNonNestedPath() && operator.IsNegativeOperator() && operator != qbtypes.FilterOperatorNotExists {
-		fieldExpr = assumeNotNull(fieldExpr)
+	// Note: here applyNotCondition will return true only if; top level path is being queried and operator is a negative operator
+	// Otherwise this code will be triggered by buildAccessNodeBranches; Where operator would've been already inverted if needed.
+	if node.IsNonNestedPath() {
+		yes, _ := applyNotCondition(operator)
+		if yes {
+			switch operator {
+			case qbtypes.FilterOperatorNotExists:
+				// skip
+			default:
+				fieldExpr = assumeNotNull(fieldExpr)
+			}
+		}
 	}
 
 	fieldExpr, formattedValue = querybuilder.DataTypeCollisionHandledFieldName(node.TerminalConfig.Key, formattedValue, fieldExpr, operator)

pkg/telemetrylogs/json_stmt_builder_test.go

@@ -220,7 +220,7 @@ func TestJSONStmtBuilder_PrimitivePaths(t *testing.T) {
 			expected: TestExpected{
 				WhereClause: "(((LOWER(toString(dynamicElement(body_v2.`user.age`, 'Int64'))) LIKE LOWER(?)) AND has(JSONAllPaths(body_v2), 'user.age')) OR ((LOWER(dynamicElement(body_v2.`user.age`, 'String')) LIKE LOWER(?)) AND has(JSONAllPaths(body_v2), 'user.age')))",
 				Args:        []any{uint64(1747945619), uint64(1747983448), "%25%", "%25%", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-				Warnings:    []string{"Key `user.age` is ambiguous, found 2 different combinations of field context / data type: [name=user.age,context=body,datatype=int64 name=user.age,context=body,datatype=string]."},
+				Warnings:    []string{"Key `user.age` is ambiguous, found 2 different combinations of field context / data type: [name=user.age,context=body,datatype=int64,jsondatatype=Int64 name=user.age,context=body,datatype=string,jsondatatype=String]."},
 			},
 		},
 		{
@@ -414,7 +414,7 @@ func TestStatementBuilderListQueryBodyPromoted(t *testing.T) {
 			},
 			expected: TestExpected{
 				Args:     []any{uint64(1747945619), uint64(1747983448), "%1.65%", 1.65, "%1.65%", 1.65, "%1.65%", 1.65, "%1.65%", 1.65, "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-				Warnings: []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64,materialized=true name=education[].parameters,context=body,datatype=[]dynamic,materialized=true]."},
+				Warnings: []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64,materialized=true,jsondatatype=Array(Nullable(Float64)) name=education[].parameters,context=body,datatype=[]dynamic,materialized=true,jsondatatype=Array(Dynamic)]."},
 			},
 			expectedErr: nil,
 		},
@@ -441,7 +441,7 @@ func TestStatementBuilderListQueryBodyPromoted(t *testing.T) {
 			},
 			expected: TestExpected{
 				Args:     []any{uint64(1747945619), uint64(1747983448), "%true%", true, "%true%", true, "%true%", true, "%true%", true, "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-				Warnings: []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64,materialized=true name=education[].parameters,context=body,datatype=[]dynamic,materialized=true]."},
+				Warnings: []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64,materialized=true,jsondatatype=Array(Nullable(Float64)) name=education[].parameters,context=body,datatype=[]dynamic,materialized=true,jsondatatype=Array(Dynamic)]."},
 			},
 			expectedErr: nil,
 		},
@@ -455,7 +455,7 @@ func TestStatementBuilderListQueryBodyPromoted(t *testing.T) {
 			},
 			expected: TestExpected{
 				Args:     []any{uint64(1747945619), uint64(1747983448), "%passed%", "passed", "%passed%", "passed", "%passed%", "passed", "%passed%", "passed", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-				Warnings: []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64,materialized=true name=education[].parameters,context=body,datatype=[]dynamic,materialized=true]."},
+				Warnings: []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64,materialized=true,jsondatatype=Array(Nullable(Float64)) name=education[].parameters,context=body,datatype=[]dynamic,materialized=true,jsondatatype=Array(Dynamic)]."},
 			},
 			expectedErr: nil,
 		},
@@ -549,7 +549,7 @@ func TestJSONStmtBuilder_ArrayPaths(t *testing.T) {
 			expected: TestExpected{
 				WhereClause: "((NOT arrayExists(`body_v2.education`-> toFloat64OrNull(dynamicElement(`body_v2.education`.`type`, 'String')) = ?, dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND (NOT arrayExists(`body_v2.education`-> dynamicElement(`body_v2.education`.`type`, 'Int64') = ?, dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))))",
 				Args:        []any{uint64(1747945619), uint64(1747983448), int64(10001), int64(10001), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-				Warnings:    []string{"Key `education[].type` is ambiguous, found 2 different combinations of field context / data type: [name=education[].type,context=body,datatype=string name=education[].type,context=body,datatype=int64]."},
+				Warnings:    []string{"Key `education[].type` is ambiguous, found 2 different combinations of field context / data type: [name=education[].type,context=body,datatype=string,jsondatatype=String name=education[].type,context=body,datatype=int64,jsondatatype=Int64]."},
 			},
 		},
 		{
@@ -576,7 +576,7 @@ func TestJSONStmtBuilder_ArrayPaths(t *testing.T) {
 			expected: TestExpected{
 				WhereClause: "(((arrayExists(`body_v2.education`-> (arrayExists(x -> LOWER(toString(x)) LIKE LOWER(?), dynamicElement(`body_v2.education`.`parameters`, 'Array(Nullable(Float64))')) OR arrayExists(x -> toFloat64(x) = ?, dynamicElement(`body_v2.education`.`parameters`, 'Array(Nullable(Float64))'))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND has(JSONAllPaths(body_v2), 'education')) OR ((arrayExists(`body_v2.education`-> (arrayExists(x -> LOWER(toString(x)) LIKE LOWER(?), arrayFilter(x->(dynamicType(x) IN ('String', 'Int64', 'Float64', 'Bool')), dynamicElement(`body_v2.education`.`parameters`, 'Array(Dynamic)'))) OR arrayExists(x -> accurateCastOrNull(x, 'Float64') = ?, arrayFilter(x->(dynamicType(x) IN ('String', 'Int64', 'Float64', 'Bool')), dynamicElement(`body_v2.education`.`parameters`, 'Array(Dynamic)')))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND has(JSONAllPaths(body_v2), 'education')))",
 				Args:        []any{uint64(1747945619), uint64(1747983448), "%1.65%", 1.65, "%1.65%", 1.65, "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-				Warnings:    []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64 name=education[].parameters,context=body,datatype=[]dynamic]."},
+				Warnings:    []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64,jsondatatype=Array(Nullable(Float64)) name=education[].parameters,context=body,datatype=[]dynamic,jsondatatype=Array(Dynamic)]."},
 			},
 		},
 		{
@@ -585,7 +585,7 @@ func TestJSONStmtBuilder_ArrayPaths(t *testing.T) {
 			expected: TestExpected{
 				WhereClause: "(((arrayExists(`body_v2.education`-> arrayExists(x -> toString(x) = ?, dynamicElement(`body_v2.education`.`parameters`, 'Array(Nullable(Float64))')), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND has(JSONAllPaths(body_v2), 'education')) OR ((arrayExists(`body_v2.education`-> arrayExists(x -> toString(x) = ?, arrayFilter(x->(dynamicType(x) IN ('String', 'Int64', 'Float64', 'Bool')), dynamicElement(`body_v2.education`.`parameters`, 'Array(Dynamic)'))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND has(JSONAllPaths(body_v2), 'education')))",
 				Args:        []any{uint64(1747945619), uint64(1747983448), "passed", "passed", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-				Warnings:    []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64 name=education[].parameters,context=body,datatype=[]dynamic]."},
+				Warnings:    []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64,jsondatatype=Array(Nullable(Float64)) name=education[].parameters,context=body,datatype=[]dynamic,jsondatatype=Array(Dynamic)]."},
 			},
 		},
 		{
@@ -594,7 +594,7 @@ func TestJSONStmtBuilder_ArrayPaths(t *testing.T) {
 			expected: TestExpected{
 				WhereClause: "(((arrayExists(`body_v2.education`-> (arrayExists(x -> LOWER(toString(x)) LIKE LOWER(?), dynamicElement(`body_v2.education`.`parameters`, 'Array(Nullable(Float64))')) OR arrayExists(x -> toString(x) = ?, dynamicElement(`body_v2.education`.`parameters`, 'Array(Nullable(Float64))'))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND has(JSONAllPaths(body_v2), 'education')) OR ((arrayExists(`body_v2.education`-> (arrayExists(x -> LOWER(toString(x)) LIKE LOWER(?), arrayFilter(x->(dynamicType(x) IN ('String', 'Int64', 'Float64', 'Bool')), dynamicElement(`body_v2.education`.`parameters`, 'Array(Dynamic)'))) OR arrayExists(x -> toString(x) = ?, arrayFilter(x->(dynamicType(x) IN ('String', 'Int64', 'Float64', 'Bool')), dynamicElement(`body_v2.education`.`parameters`, 'Array(Dynamic)')))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND has(JSONAllPaths(body_v2), 'education')))",
 				Args:        []any{uint64(1747945619), uint64(1747983448), "%passed%", "passed", "%passed%", "passed", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-				Warnings:    []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64 name=education[].parameters,context=body,datatype=[]dynamic]."},
+				Warnings:    []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64,jsondatatype=Array(Nullable(Float64)) name=education[].parameters,context=body,datatype=[]dynamic,jsondatatype=Array(Dynamic)]."},
 			},
 		},
 		{
@@ -603,7 +603,7 @@ func TestJSONStmtBuilder_ArrayPaths(t *testing.T) {
 			expected: TestExpected{
 				WhereClause: "(((arrayExists(`body_v2.education`-> arrayExists(x -> toFloat64(x) IN (?, ?), dynamicElement(`body_v2.education`.`parameters`, 'Array(Nullable(Float64))')), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND has(JSONAllPaths(body_v2), 'education')) OR ((arrayExists(`body_v2.education`-> arrayExists(x -> toString(x) IN (?, ?), arrayFilter(x->(dynamicType(x) IN ('String', 'Int64', 'Float64', 'Bool')), dynamicElement(`body_v2.education`.`parameters`, 'Array(Dynamic)'))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND has(JSONAllPaths(body_v2), 'education')))",
 				Args:        []any{uint64(1747945619), uint64(1747983448), 1.65, 1.99, "1.65", "1.99", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-				Warnings:    []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64 name=education[].parameters,context=body,datatype=[]dynamic]."},
+				Warnings:    []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64,jsondatatype=Array(Nullable(Float64)) name=education[].parameters,context=body,datatype=[]dynamic,jsondatatype=Array(Dynamic)]."},
 			},
 		},
 		{
@@ -612,7 +612,7 @@ func TestJSONStmtBuilder_ArrayPaths(t *testing.T) {
 			expected: TestExpected{
 				WhereClause: "((NOT arrayExists(`body_v2.education`-> (arrayExists(x -> LOWER(toString(x)) LIKE LOWER(?), dynamicElement(`body_v2.education`.`parameters`, 'Array(Nullable(Float64))')) OR arrayExists(x -> toFloat64(x) = ?, dynamicElement(`body_v2.education`.`parameters`, 'Array(Nullable(Float64))'))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND (NOT arrayExists(`body_v2.education`-> (arrayExists(x -> LOWER(toString(x)) LIKE LOWER(?), arrayFilter(x->(dynamicType(x) IN ('String', 'Int64', 'Float64', 'Bool')), dynamicElement(`body_v2.education`.`parameters`, 'Array(Dynamic)'))) OR arrayExists(x -> accurateCastOrNull(x, 'Float64') = ?, arrayFilter(x->(dynamicType(x) IN ('String', 'Int64', 'Float64', 'Bool')), dynamicElement(`body_v2.education`.`parameters`, 'Array(Dynamic)')))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))))",
 				Args:        []any{uint64(1747945619), uint64(1747983448), "%1.65%", float64(1.65), "%1.65%", float64(1.65), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-				Warnings:    []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64 name=education[].parameters,context=body,datatype=[]dynamic]."},
+				Warnings:    []string{"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64,jsondatatype=Array(Nullable(Float64)) name=education[].parameters,context=body,datatype=[]dynamic,jsondatatype=Array(Dynamic)]."},
 			},
 		},
 		{
@@ -622,7 +622,7 @@ func TestJSONStmtBuilder_ArrayPaths(t *testing.T) {
 				WhereClause: "(has(arrayFlatten(arrayConcat(arrayMap(`body_v2.education`->dynamicElement(`body_v2.education`.`parameters`, 'Array(Nullable(Float64))'), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))')))), ?) OR has(arrayFlatten(arrayConcat(arrayMap(`body_v2.education`->dynamicElement(`body_v2.education`.`parameters`, 'Array(Dynamic)'), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))')))), ?))",
 				Args:        []any{uint64(1747945619), uint64(1747983448), 1.65, 1.65, "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
 				Warnings: []string{
-					"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64 name=education[].parameters,context=body,datatype=[]dynamic].",
+					"Key `education[].parameters` is ambiguous, found 2 different combinations of field context / data type: [name=education[].parameters,context=body,datatype=[]float64,jsondatatype=Array(Nullable(Float64)) name=education[].parameters,context=body,datatype=[]dynamic,jsondatatype=Array(Dynamic)].",
 				},
 			},
 		},
@@ -702,7 +702,7 @@ func TestJSONStmtBuilder_ArrayPaths(t *testing.T) {
 			expected: TestExpected{
 				WhereClause: "(((arrayExists(`body_v2.interests`-> arrayExists(`body_v2.interests[].entities`-> arrayExists(`body_v2.interests[].entities[].reviews`-> arrayExists(`body_v2.interests[].entities[].reviews[].entries`-> arrayExists(`body_v2.interests[].entities[].reviews[].entries[].metadata`-> arrayExists(`body_v2.interests[].entities[].reviews[].entries[].metadata[].positions`-> (arrayExists(x -> LOWER(toString(x)) LIKE LOWER(?), dynamicElement(`body_v2.interests[].entities[].reviews[].entries[].metadata[].positions`.`ratings`, 'Array(Nullable(Int64))')) OR arrayExists(x -> toFloat64(x) = ?, dynamicElement(`body_v2.interests[].entities[].reviews[].entries[].metadata[].positions`.`ratings`, 'Array(Nullable(Int64))'))), dynamicElement(`body_v2.interests[].entities[].reviews[].entries[].metadata`.`positions`, 'Array(JSON(max_dynamic_types=0, max_dynamic_paths=0))')), dynamicElement(`body_v2.interests[].entities[].reviews[].entries`.`metadata`, 'Array(JSON(max_dynamic_types=1, max_dynamic_paths=0))')), dynamicElement(`body_v2.interests[].entities[].reviews`.`entries`, 'Array(JSON(max_dynamic_types=2, max_dynamic_paths=0))')), dynamicElement(`body_v2.interests[].entities`.`reviews`, 'Array(JSON(max_dynamic_types=4, max_dynamic_paths=0))')), dynamicElement(`body_v2.interests`.`entities`, 'Array(JSON(max_dynamic_types=8, max_dynamic_paths=0))')), dynamicElement(body_v2.`interests`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND has(JSONAllPaths(body_v2), 'interests')) OR ((arrayExists(`body_v2.interests`-> arrayExists(`body_v2.interests[].entities`-> arrayExists(`body_v2.interests[].entities[].reviews`-> arrayExists(`body_v2.interests[].entities[].reviews[].entries`-> arrayExists(`body_v2.interests[].entities[].reviews[].entries[].metadata`-> arrayExists(`body_v2.interests[].entities[].reviews[].entries[].metadata[].positions`-> (arrayExists(x -> LOWER(x) LIKE LOWER(?), dynamicElement(`body_v2.interests[].entities[].reviews[].entries[].metadata[].positions`.`ratings`, 'Array(Nullable(String))')) OR arrayExists(x -> toFloat64OrNull(x) = ?, dynamicElement(`body_v2.interests[].entities[].reviews[].entries[].metadata[].positions`.`ratings`, 'Array(Nullable(String))'))), dynamicElement(`body_v2.interests[].entities[].reviews[].entries[].metadata`.`positions`, 'Array(JSON(max_dynamic_types=0, max_dynamic_paths=0))')), dynamicElement(`body_v2.interests[].entities[].reviews[].entries`.`metadata`, 'Array(JSON(max_dynamic_types=1, max_dynamic_paths=0))')), dynamicElement(`body_v2.interests[].entities[].reviews`.`entries`, 'Array(JSON(max_dynamic_types=2, max_dynamic_paths=0))')), dynamicElement(`body_v2.interests[].entities`.`reviews`, 'Array(JSON(max_dynamic_types=4, max_dynamic_paths=0))')), dynamicElement(`body_v2.interests`.`entities`, 'Array(JSON(max_dynamic_types=8, max_dynamic_paths=0))')), dynamicElement(body_v2.`interests`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))'))) AND has(JSONAllPaths(body_v2), 'interests')))",
 				Args:        []any{uint64(1747945619), uint64(1747983448), "%4%", float64(4), "%4%", float64(4), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-				Warnings:    []string{"Key `interests[].entities[].reviews[].entries[].metadata[].positions[].ratings` is ambiguous, found 2 different combinations of field context / data type: [name=interests[].entities[].reviews[].entries[].metadata[].positions[].ratings,context=body,datatype=[]int64 name=interests[].entities[].reviews[].entries[].metadata[].positions[].ratings,context=body,datatype=[]string]."},
+				Warnings:    []string{"Key `interests[].entities[].reviews[].entries[].metadata[].positions[].ratings` is ambiguous, found 2 different combinations of field context / data type: [name=interests[].entities[].reviews[].entries[].metadata[].positions[].ratings,context=body,datatype=[]int64,jsondatatype=Array(Nullable(Int64)) name=interests[].entities[].reviews[].entries[].metadata[].positions[].ratings,context=body,datatype=[]string,jsondatatype=Array(Nullable(String))]."},
 			},
 		},
 		{
@@ -844,10 +844,7 @@ func TestJSONStmtBuilder_IndexedPaths(t *testing.T) {
 				Args:        []any{uint64(1747945619), uint64(1747983448), float64(110001), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
 			},
 		},
-		// ── indexed exists: index is skipped; emits plain IS NOT NULL ───────────
-		// EXISTS/NOT EXISTS bypass the index because indexed columns store a default
-		// empty value for absent fields, making != "" unreliable for existence checks
-		// (a field with value "" would be incorrectly excluded).
+		// ── indexed exists: emits assumeNotNull != nil AND dynamicElement IS NOT NULL ─
 		{
 			name: "Indexed String Exists",
 			query: qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]{
@@ -856,8 +853,8 @@ func TestJSONStmtBuilder_IndexedPaths(t *testing.T) {
 				Limit:  10,
 			},
 			expected: TestExpected{
-				WhereClause: "(dynamicElement(body_v2.`user.name`, 'String') IS NOT NULL)",
-				Args:        []any{uint64(1747945619), uint64(1747983448), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
+				WhereClause: "((assumeNotNull(dynamicElement(body_v2.`user.name`, 'String')) <> ? AND dynamicElement(body_v2.`user.name`, 'String') IS NOT NULL))",
+				Args:        []any{uint64(1747945619), uint64(1747983448), "", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
 			},
 		},
 		// ── indexed not-equal: assumeNotNull wrapping + no path index ─────────
@@ -873,6 +870,22 @@ func TestJSONStmtBuilder_IndexedPaths(t *testing.T) {
 				Args:        []any{uint64(1747945619), uint64(1747983448), "alice", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
 			},
 		},
+		// ── indexed not-exists: assumeNotNull = "" AND assumeNotNull IS NOT NULL ─
+		// FilterOperatorNotExists → Equal + emptyValue("") in the indexed branch,
+		// then a second condition flipped to Exists (IS NOT NULL) on the same
+		// assumeNotNull expr, producing AND(= "", IS NOT NULL).
+		{
+			name: "Indexed String NotExists",
+			query: qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]{
+				Signal: telemetrytypes.SignalLogs,
+				Filter: &qbtypes.Filter{Expression: "body.user.name NOT EXISTS"},
+				Limit:  10,
+			},
+			expected: TestExpected{
+				WhereClause: "((assumeNotNull(dynamicElement(body_v2.`user.name`, 'String')) = ? AND dynamicElement(body_v2.`user.name`, 'String') IS NULL))",
+				Args:        []any{uint64(1747945619), uint64(1747983448), "", "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
+			},
+		},
 
 		// ── special-character indexed paths ───────────────────────────────────
 		{
@@ -929,171 +942,20 @@ func TestJSONStmtBuilder_IndexedPaths(t *testing.T) {
 		})
 	}
 }
-
-func TestJSONStmtBuilder_SelectField(t *testing.T) {
-	enable, disable := jsonQueryTestUtil(t)
-	enable()
-	defer disable()
-	statementBuilder := buildJSONTestStatementBuilder(t, false)
-
-	cases := []struct {
-		name                string
-		requestType         qbtypes.RequestType
-		query               qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]
-		expected            qbtypes.Statement
-		expectedErrContains string
-	}{
-		{
-			name:        "select_x_education[].awards[].participated[].members",
-			requestType: qbtypes.RequestTypeRaw,
-			query: qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]{
-				Signal:       telemetrytypes.SignalLogs,
-				StepInterval: qbtypes.Step{Duration: 30 * time.Second},
-				Limit:        10,
-				SelectFields: []telemetrytypes.TelemetryFieldKey{
-					{
-						Name: "education[].awards[].participated[].members",
-					},
-				},
-				Filter: &qbtypes.Filter{
-					Expression: "user.name exists",
-				},
-			},
-			expected: qbtypes.Statement{
-				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE true AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, arrayFlatten(arrayConcat(arrayMap(`body_v2.education`->arrayConcat(arrayMap(`body_v2.education[].awards`->arrayConcat(arrayMap(`body_v2.education[].awards[].participated`->dynamicElement(`body_v2.education[].awards[].participated`.`members`, 'Array(Nullable(String))'), dynamicElement(`body_v2.education[].awards`.`participated`, 'Array(JSON(max_dynamic_types=4, max_dynamic_paths=0))')), arrayMap(`body_v2.education[].awards[].participated`->dynamicElement(`body_v2.education[].awards[].participated`.`members`, 'Array(Nullable(String))'), arrayMap(x->assumeNotNull(dynamicElement(x, 'JSON')), arrayFilter(x->(dynamicType(x) = 'JSON'), dynamicElement(`body_v2.education[].awards`.`participated`, 'Array(Dynamic)'))))), dynamicElement(`body_v2.education`.`awards`, 'Array(JSON(max_dynamic_types=8, max_dynamic_paths=0))')), arrayMap(`body_v2.education[].awards`->arrayConcat(arrayMap(`body_v2.education[].awards[].participated`->dynamicElement(`body_v2.education[].awards[].participated`.`members`, 'Array(Nullable(String))'), dynamicElement(`body_v2.education[].awards`.`participated`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=256))')), arrayMap(`body_v2.education[].awards[].participated`->dynamicElement(`body_v2.education[].awards[].participated`.`members`, 'Array(Nullable(String))'), arrayMap(x->assumeNotNull(dynamicElement(x, 'JSON')), arrayFilter(x->(dynamicType(x) = 'JSON'), dynamicElement(`body_v2.education[].awards`.`participated`, 'Array(Dynamic)'))))), arrayMap(x->assumeNotNull(dynamicElement(x, 'JSON')), arrayFilter(x->(dynamicType(x) = 'JSON'), dynamicElement(`body_v2.education`.`awards`, 'Array(Dynamic)'))))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))')))) AS `education[].awards[].participated[].members` FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND (dynamicElement(body_v2.`user.name`, 'String') IS NOT NULL) AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
-				Args:  []any{uint64(1747945619), uint64(1747983448), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-			},
-		},
-		{
-			name:        "select_x_user.name",
-			requestType: qbtypes.RequestTypeRaw,
-			query: qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]{
-				Signal:       telemetrytypes.SignalLogs,
-				StepInterval: qbtypes.Step{Duration: 30 * time.Second},
-				Limit:        10,
-				SelectFields: []telemetrytypes.TelemetryFieldKey{
-					{
-						Name: "user.name",
-					},
-				},
-			},
-			expected: qbtypes.Statement{
-				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, dynamicElement(body_v2.`user.name`, 'String') AS `user.name` FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? LIMIT ?",
-				Args:  []any{uint64(1747945619), uint64(1747983448), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-			},
-		},
-	}
-
-	for _, c := range cases {
-		t.Run(c.name, func(t *testing.T) {
-			q, err := statementBuilder.Build(context.Background(), 1747947419000, 1747983448000, c.requestType, c.query, nil)
-			if c.expectedErrContains != "" {
-				require.Error(t, err)
-				require.Contains(t, err.Error(), c.expectedErrContains)
-			} else {
-				require.NoError(t, err)
-				require.Equal(t, c.expected.Query, q.Query)
-				require.Equal(t, c.expected.Args, q.Args)
-				require.Equal(t, c.expected.Warnings, q.Warnings)
-			}
-		})
-	}
-}
-
-func TestJSONStmtBuilder_OrderBy(t *testing.T) {
-	enable, disable := jsonQueryTestUtil(t)
-	enable()
-	defer disable()
-	statementBuilder := buildJSONTestStatementBuilder(t, false)
-
-	cases := []struct {
-		name                string
-		requestType         qbtypes.RequestType
-		query               qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]
-		expected            qbtypes.Statement
-		expectedErrContains string
-	}{
-		{
-			name:        "order_by_education[].awards[].participated[].members",
-			requestType: qbtypes.RequestTypeRaw,
-			query: qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]{
-				Signal:       telemetrytypes.SignalLogs,
-				StepInterval: qbtypes.Step{Duration: 30 * time.Second},
-				Limit:        10,
-				Order: []qbtypes.OrderBy{
-					{
-						Key: qbtypes.OrderByKey{
-							TelemetryFieldKey: telemetrytypes.TelemetryFieldKey{
-								Name: "education[].awards[].participated[].members",
-							},
-						},
-						Direction: qbtypes.OrderDirectionAsc,
-					},
-				},
-				Filter: &qbtypes.Filter{
-					Expression: "user.name exists",
-				},
-			},
-			expected: qbtypes.Statement{
-				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE true AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND (dynamicElement(body_v2.`user.name`, 'String') IS NOT NULL) AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? ORDER BY arrayFlatten(arrayConcat(arrayMap(`body_v2.education`->arrayConcat(arrayMap(`body_v2.education[].awards`->arrayConcat(arrayMap(`body_v2.education[].awards[].participated`->dynamicElement(`body_v2.education[].awards[].participated`.`members`, 'Array(Nullable(String))'), dynamicElement(`body_v2.education[].awards`.`participated`, 'Array(JSON(max_dynamic_types=4, max_dynamic_paths=0))')), arrayMap(`body_v2.education[].awards[].participated`->dynamicElement(`body_v2.education[].awards[].participated`.`members`, 'Array(Nullable(String))'), arrayMap(x->assumeNotNull(dynamicElement(x, 'JSON')), arrayFilter(x->(dynamicType(x) = 'JSON'), dynamicElement(`body_v2.education[].awards`.`participated`, 'Array(Dynamic)'))))), dynamicElement(`body_v2.education`.`awards`, 'Array(JSON(max_dynamic_types=8, max_dynamic_paths=0))')), arrayMap(`body_v2.education[].awards`->arrayConcat(arrayMap(`body_v2.education[].awards[].participated`->dynamicElement(`body_v2.education[].awards[].participated`.`members`, 'Array(Nullable(String))'), dynamicElement(`body_v2.education[].awards`.`participated`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=256))')), arrayMap(`body_v2.education[].awards[].participated`->dynamicElement(`body_v2.education[].awards[].participated`.`members`, 'Array(Nullable(String))'), arrayMap(x->assumeNotNull(dynamicElement(x, 'JSON')), arrayFilter(x->(dynamicType(x) = 'JSON'), dynamicElement(`body_v2.education[].awards`.`participated`, 'Array(Dynamic)'))))), arrayMap(x->assumeNotNull(dynamicElement(x, 'JSON')), arrayFilter(x->(dynamicType(x) = 'JSON'), dynamicElement(`body_v2.education`.`awards`, 'Array(Dynamic)'))))), dynamicElement(body_v2.`education`, 'Array(JSON(max_dynamic_types=16, max_dynamic_paths=0))')))) AS `education[].awards[].participated[].members` asc LIMIT ?",
-				Args:  []any{uint64(1747945619), uint64(1747983448), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-			},
-		},
-		{
-			name:        "order_by_user.name",
-			requestType: qbtypes.RequestTypeRaw,
-			query: qbtypes.QueryBuilderQuery[qbtypes.LogAggregation]{
-				Signal:       telemetrytypes.SignalLogs,
-				StepInterval: qbtypes.Step{Duration: 30 * time.Second},
-				Limit:        10,
-				Order: []qbtypes.OrderBy{
-					{
-						Key: qbtypes.OrderByKey{
-							TelemetryFieldKey: telemetrytypes.TelemetryFieldKey{
-								Name: "user.name",
-							},
-						},
-						Direction: qbtypes.OrderDirectionAsc,
-					},
-				},
-			},
-			expected: qbtypes.Statement{
-				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_logs.distributed_logs_v2_resource WHERE seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp, id, trace_id, span_id, trace_flags, severity_text, severity_number, scope_name, scope_version, body_v2 as body, attributes_string, attributes_number, attributes_bool, resources_string, scope_string FROM signoz_logs.distributed_logs_v2 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND timestamp >= ? AND ts_bucket_start >= ? AND timestamp < ? AND ts_bucket_start <= ? ORDER BY dynamicElement(body_v2.`user.name`, 'String') AS `user.name` asc LIMIT ?",
-				Args:  []any{uint64(1747945619), uint64(1747983448), "1747947419000000000", uint64(1747945619), "1747983448000000000", uint64(1747983448), 10},
-			},
-		},
-	}
-
-	for _, c := range cases {
-		t.Run(c.name, func(t *testing.T) {
-			q, err := statementBuilder.Build(context.Background(), 1747947419000, 1747983448000, c.requestType, c.query, nil)
-			if c.expectedErrContains != "" {
-				require.Error(t, err)
-				require.Contains(t, err.Error(), c.expectedErrContains)
-			} else {
-				require.NoError(t, err)
-				require.Equal(t, c.expected.Query, q.Query)
-				require.Equal(t, c.expected.Args, q.Args)
-				require.Equal(t, c.expected.Warnings, q.Warnings)
-			}
-		})
-	}
-}
-
 func buildTestTelemetryMetadataStore(t *testing.T, addIndexes bool) *telemetrytypestest.MockMetadataStore {
 	mockMetadataStore := telemetrytypestest.NewMockMetadataStore()
 	mockMetadataStore.SetStaticFields(IntrinsicFields)
 	types, _ := telemetrytypes.TestJSONTypeSet()
-	for path, fieldDataTypes := range types {
-		for _, fdt := range fieldDataTypes {
+	for path, jsonTypes := range types {
+		for _, jsonType := range jsonTypes {
 			key := &telemetrytypes.TelemetryFieldKey{
 				Name:          path,
 				Signal:        telemetrytypes.SignalLogs,
 				FieldContext:  telemetrytypes.FieldContextBody,
-				FieldDataType: fdt,
+				FieldDataType: telemetrytypes.MappingJSONDataTypeToFieldDataType[jsonType],
+				JSONDataType:  &jsonType,
 			}
 			if addIndexes {
-				jsonType := telemetrytypes.MappingFieldDataTypeToJSONDataType[fdt]
 				idx := slices.IndexFunc(telemetrytypes.TestIndexedPaths, func(entry telemetrytypes.TestIndexedPathEntry) bool {
 					return entry.Path == path && entry.Type == jsonType
 				})

pkg/telemetrylogs/stmt_builder_test.go

@@ -875,6 +875,7 @@ func TestAdjustKey(t *testing.T) {
 			require.Equal(t, c.expectedKey.FieldContext, key.FieldContext, "field context should match")
 			require.Equal(t, c.expectedKey.FieldDataType, key.FieldDataType, "field data type should match")
 			require.Equal(t, c.expectedKey.Materialized, key.Materialized, "materialized should match")
+			require.Equal(t, c.expectedKey.JSONDataType, key.JSONDataType, "json data type should match")
 			require.Equal(t, c.expectedKey.Indexes, key.Indexes, "json exists should match")
 		})
 	}

pkg/telemetrymetadata/body_json_metadata.go

@@ -21,84 +21,133 @@ import (
 )
 
 var (
+	defaultPathLimit = 100 // Default limit to prevent full table scans
+
+	CodeUnknownJSONDataType     = errors.MustNewCode("unknown_json_data_type")
 	CodeFailLoadPromotedPaths   = errors.MustNewCode("fail_load_promoted_paths")
 	CodeFailCheckPathPromoted   = errors.MustNewCode("fail_check_path_promoted")
+	CodeFailIterateBodyJSONKeys = errors.MustNewCode("fail_iterate_body_json_keys")
+	CodeFailExtractBodyJSONKeys = errors.MustNewCode("fail_extract_body_json_keys")
 	CodeFailLoadLogsJSONIndexes = errors.MustNewCode("fail_load_logs_json_indexes")
 	CodeFailListJSONValues      = errors.MustNewCode("fail_list_json_values")
 	CodeFailScanJSONValue       = errors.MustNewCode("fail_scan_json_value")
 	CodeFailScanVariant         = errors.MustNewCode("fail_scan_variant")
+	CodeFailBuildJSONPathsQuery = errors.MustNewCode("fail_build_json_paths_query")
+	CodeNoPathsToQueryIndexes   = errors.MustNewCode("no_paths_to_query_indexes_provided")
 
 	CodeFailedToPrepareBatch = errors.MustNewCode("failed_to_prepare_batch_promoted_paths")
 	CodeFailedToSendBatch    = errors.MustNewCode("failed_to_send_batch_promoted_paths")
 	CodeFailedToAppendPath   = errors.MustNewCode("failed_to_append_path_promoted_paths")
 )
 
-// enrichJSONKeys enriches body-context keys with promoted path info, indexes,
-// and JSON access plans. parentTypeCache contains parent array types (ArrayJSON/ArrayDynamic)
-// pre-fetched in the main UNION query.
+// fetchBodyJSONPaths extracts body JSON paths from the path_types table
+// This function can be used by both JSONQueryBuilder and metadata extraction
+// uniquePathLimit: 0 for no limit, >0 for maximum number of unique paths to return
+//   - For startup load: set to 10000 to get top 10k unique paths
+//   - For lookup: set to 0 (no limit needed for single path)
+//   - For metadata API: set to desired pagination limit
 //
-// NOTE: enrichment can not work with FuzzySelectors; QB requests exact matches for query building so
-// parentTypeCache will actually have proper matches and
-// FuzzyMatching is for Suggestions API so enrichment is not needed
-func (t *telemetryMetaStore) enrichJSONKeys(ctx context.Context, selectors []*telemetrytypes.FieldKeySelector, keys []*telemetrytypes.TelemetryFieldKey, parentTypeCache map[string][]telemetrytypes.FieldDataType) error {
-	mapOfExactSelectors := make(map[string]*telemetrytypes.FieldKeySelector)
-	for _, selector := range selectors {
-		if selector.SelectorMatchType != telemetrytypes.FieldSelectorMatchTypeExact {
-			continue
+// searchOperator: LIKE for pattern matching, EQUAL for exact match.
+func (t *telemetryMetaStore) fetchBodyJSONPaths(ctx context.Context,
+	fieldKeySelectors []*telemetrytypes.FieldKeySelector) ([]*telemetrytypes.TelemetryFieldKey, []string, bool, error) {
+	ctx = ctxtypes.NewContextWithCommentVals(ctx, map[string]string{
+		instrumentationtypes.TelemetrySignal:  telemetrytypes.SignalLogs.StringValue(),
+		instrumentationtypes.CodeNamespace:    "metadata",
+		instrumentationtypes.CodeFunctionName: "fetchBodyJSONPaths",
+	})
+
+	query, args, limit := buildGetBodyJSONPathsQuery(fieldKeySelectors)
+	rows, err := t.telemetrystore.ClickhouseDB().Query(ctx, query, args...)
+	if err != nil {
+		return nil, nil, false, errors.WrapInternalf(err, CodeFailExtractBodyJSONKeys, "failed to extract body JSON keys")
+	}
+	defer rows.Close()
+
+	fieldKeys := []*telemetrytypes.TelemetryFieldKey{}
+	paths := []string{}
+	rowCount := 0
+	for rows.Next() {
+		var path string
+		var typesArray []string // ClickHouse returns array as []string
+		var lastSeen uint64
+
+		err = rows.Scan(&path, &typesArray, &lastSeen)
+		if err != nil {
+			return nil, nil, false, errors.WrapInternalf(err, CodeFailExtractBodyJSONKeys, "failed to scan body JSON key row")
 		}
 
-		mapOfExactSelectors[selector.Name] = selector
-	}
-
-	var filteredKeys []*telemetrytypes.TelemetryFieldKey
-	for _, key := range keys {
-		if key.FieldContext == telemetrytypes.FieldContextBody && mapOfExactSelectors[key.Name] != nil {
-			filteredKeys = append(filteredKeys, key)
+		for _, typ := range typesArray {
+			mapping, found := telemetrytypes.MappingStringToJSONDataType[typ]
+			if !found {
+				t.logger.ErrorContext(ctx, "failed to map type string to JSON data type", slog.String("type", typ), slog.String("path", path))
+				continue
+			}
+			fieldKeys = append(fieldKeys, &telemetrytypes.TelemetryFieldKey{
+				Name:          path,
+				Signal:        telemetrytypes.SignalLogs,
+				FieldContext:  telemetrytypes.FieldContextBody,
+				FieldDataType: telemetrytypes.MappingJSONDataTypeToFieldDataType[mapping],
+				JSONDataType:  &mapping,
+			})
 		}
+
+		paths = append(paths, path)
+		rowCount++
+	}
+	if rows.Err() != nil {
+		return nil, nil, false, errors.WrapInternalf(rows.Err(), CodeFailIterateBodyJSONKeys, "error iterating body JSON keys")
 	}
 
-	if len(filteredKeys) == 0 {
-		return nil
-	}
-
-	paths := make([]string, 0, len(filteredKeys))
-	for _, key := range filteredKeys {
-		paths = append(paths, key.Name)
-	}
-
-	// fetch promoted paths
-	promoted, err := t.GetPromotedPaths(ctx, paths...)
-	if err != nil {
-		return err
-	}
-
-	// fetch JSON path indexes
-	indexes, err := t.getJSONPathIndexes(ctx, paths...)
-	if err != nil {
-		return err
-	}
-
-	// apply promoted/index metadata to keys
-	for _, key := range filteredKeys {
-		promotedKey := strings.Split(key.Name, telemetrytypes.ArraySep)[0]
-		key.Materialized = promoted[promotedKey]
-		key.Indexes = indexes[key.Name]
-	}
-
-	// build JSON access plans using the pre-fetched parent type cache
-	return t.buildJSONPlans(ctx, filteredKeys, parentTypeCache)
+	return fieldKeys, paths, rowCount <= limit, nil
 }
 
-// buildJSONPlans builds JSON access plans for the given keys
-// using the provided parent type cache (pre-fetched in the main UNION query).
-func (t *telemetryMetaStore) buildJSONPlans(_ context.Context, keys []*telemetrytypes.TelemetryFieldKey, typeCache map[string][]telemetrytypes.FieldDataType) error {
-	if len(keys) == 0 {
-		return nil
+func (t *telemetryMetaStore) buildBodyJSONPaths(ctx context.Context,
+	fieldKeySelectors []*telemetrytypes.FieldKeySelector) ([]*telemetrytypes.TelemetryFieldKey, bool, error) {
+
+	fieldKeys, paths, finished, err := t.fetchBodyJSONPaths(ctx, fieldKeySelectors)
+	if err != nil {
+		return nil, false, err
 	}
 
-	columnMeta := t.jsonColumnMetadata[telemetrytypes.SignalLogs][telemetrytypes.FieldContextBody]
+	promoted, err := t.GetPromotedPaths(ctx, paths...)
+	if err != nil {
+		return nil, false, err
+	}
+
+	indexes, err := t.getJSONPathIndexes(ctx, paths...)
+	if err != nil {
+		return nil, false, err
+	}
+
+	for _, fieldKey := range fieldKeys {
+		promotedKey := strings.Split(fieldKey.Name, telemetrytypes.ArraySep)[0]
+		fieldKey.Materialized = promoted[promotedKey]
+		fieldKey.Indexes = indexes[fieldKey.Name]
+	}
+
+	return fieldKeys, finished, t.buildJSONPlans(ctx, fieldKeys)
+}
+
+func (t *telemetryMetaStore) buildJSONPlans(ctx context.Context, keys []*telemetrytypes.TelemetryFieldKey) error {
+	parentSelectors := make([]*telemetrytypes.FieldKeySelector, 0, len(keys))
 	for _, key := range keys {
-		if err := key.SetJSONAccessPlan(columnMeta, typeCache); err != nil {
+		parentSelectors = append(parentSelectors, key.ArrayParentSelectors()...)
+	}
+
+	parentKeys, _, _, err := t.fetchBodyJSONPaths(ctx, parentSelectors)
+	if err != nil {
+		return err
+	}
+
+	typeCache := make(map[string][]telemetrytypes.JSONDataType)
+	for _, key := range parentKeys {
+		typeCache[key.Name] = append(typeCache[key.Name], *key.JSONDataType)
+	}
+
+	// build plans for keys now
+	for _, key := range keys {
+		err = key.SetJSONAccessPlan(t.jsonColumnMetadata[telemetrytypes.SignalLogs][telemetrytypes.FieldContextBody], typeCache)
+		if err != nil {
 			return err
 		}
 	}
@@ -106,6 +155,51 @@ func (t *telemetryMetaStore) buildJSONPlans(_ context.Context, keys []*telemetry
 	return nil
 }
 
+func buildGetBodyJSONPathsQuery(fieldKeySelectors []*telemetrytypes.FieldKeySelector) (string, []any, int) {
+	if len(fieldKeySelectors) == 0 {
+		return "", nil, defaultPathLimit
+	}
+	from := fmt.Sprintf("%s.%s", DBName, PathTypesTableName)
+
+	// Build a better query using GROUP BY to deduplicate at database level
+	// This aggregates all types per path and gets the max last_seen, then applies LIMIT
+	sb := sqlbuilder.Select(
+		"path",
+		"groupArray(DISTINCT type) AS types",
+		"max(last_seen) AS last_seen",
+	).From(from)
+
+	limit := 0
+	// Add search filter if provided
+	orClauses := []string{}
+	for _, fieldKeySelector := range fieldKeySelectors {
+		// replace [*] with []
+		fieldKeySelector.Name = strings.ReplaceAll(fieldKeySelector.Name, telemetrytypes.ArrayAnyIndex, telemetrytypes.ArraySep)
+		// Extract search text for body JSON keys
+		keyName := CleanPathPrefixes(fieldKeySelector.Name)
+		if fieldKeySelector.SelectorMatchType == telemetrytypes.FieldSelectorMatchTypeExact {
+			orClauses = append(orClauses, sb.Equal("path", keyName))
+		} else {
+			// Pattern matching for metadata API (defaults to LIKE behavior for other operators)
+			orClauses = append(orClauses, sb.ILike("path", fmt.Sprintf("%%%s%%", querybuilder.FormatValueForContains(keyName))))
+		}
+		limit += fieldKeySelector.Limit
+	}
+	sb.Where(sb.Or(orClauses...))
+	// Group by path to get unique paths with aggregated types
+	sb.GroupBy("path")
+
+	// Order by max last_seen to get most recent paths first
+	sb.OrderBy("last_seen DESC")
+	if limit == 0 {
+		limit = defaultPathLimit
+	}
+	sb.Limit(limit)
+
+	query, args := sb.BuildWithFlavor(sqlbuilder.ClickHouse)
+	return query, args, limit
+}
+
 func (t *telemetryMetaStore) getJSONPathIndexes(ctx context.Context, paths ...string) (map[string][]telemetrytypes.JSONDataTypeIndex, error) {
 	filteredPaths := []string{}
 	for _, path := range paths {

pkg/telemetrymetadata/body_json_metadata_test.go

@@ -7,9 +7,99 @@ import (
 	"github.com/SigNoz/signoz-otel-collector/constants"
 	"github.com/SigNoz/signoz/pkg/querybuilder"
 	"github.com/SigNoz/signoz/pkg/telemetrylogs"
+	"github.com/SigNoz/signoz/pkg/types/telemetrytypes"
 	"github.com/stretchr/testify/require"
 )
 
+func TestBuildGetBodyJSONPathsQuery(t *testing.T) {
+	testCases := []struct {
+		name              string
+		fieldKeySelectors []*telemetrytypes.FieldKeySelector
+		expectedSQL       string
+		expectedArgs      []any
+		expectedLimit     int
+	}{
+
+		{
+			name: "Single search text with EQUAL operator",
+			fieldKeySelectors: []*telemetrytypes.FieldKeySelector{
+				{
+					Name:              "user.name",
+					SelectorMatchType: telemetrytypes.FieldSelectorMatchTypeExact,
+				},
+			},
+			expectedSQL:   "SELECT path, groupArray(DISTINCT type) AS types, max(last_seen) AS last_seen FROM signoz_metadata.distributed_field_keys WHERE (path = ?) GROUP BY path ORDER BY last_seen DESC LIMIT ?",
+			expectedArgs:  []any{"user.name", defaultPathLimit},
+			expectedLimit: defaultPathLimit,
+		},
+		{
+			name: "Single search text with LIKE operator",
+			fieldKeySelectors: []*telemetrytypes.FieldKeySelector{
+				{
+					Name:              "user",
+					SelectorMatchType: telemetrytypes.FieldSelectorMatchTypeFuzzy,
+				},
+			},
+			expectedSQL:   "SELECT path, groupArray(DISTINCT type) AS types, max(last_seen) AS last_seen FROM signoz_metadata.distributed_field_keys WHERE (LOWER(path) LIKE LOWER(?)) GROUP BY path ORDER BY last_seen DESC LIMIT ?",
+			expectedArgs:  []any{"%user%", 100},
+			expectedLimit: 100,
+		},
+		{
+			name: "Multiple search texts with EQUAL operator",
+			fieldKeySelectors: []*telemetrytypes.FieldKeySelector{
+				{
+					Name:              "user.name",
+					SelectorMatchType: telemetrytypes.FieldSelectorMatchTypeExact,
+				},
+				{
+					Name:              "user.age",
+					SelectorMatchType: telemetrytypes.FieldSelectorMatchTypeExact,
+				},
+			},
+			expectedSQL:   "SELECT path, groupArray(DISTINCT type) AS types, max(last_seen) AS last_seen FROM signoz_metadata.distributed_field_keys WHERE (path = ? OR path = ?) GROUP BY path ORDER BY last_seen DESC LIMIT ?",
+			expectedArgs:  []any{"user.name", "user.age", defaultPathLimit},
+			expectedLimit: defaultPathLimit,
+		},
+		{
+			name: "Multiple search texts with LIKE operator",
+			fieldKeySelectors: []*telemetrytypes.FieldKeySelector{
+				{
+					Name:              "user",
+					SelectorMatchType: telemetrytypes.FieldSelectorMatchTypeFuzzy,
+				},
+				{
+					Name:              "admin",
+					SelectorMatchType: telemetrytypes.FieldSelectorMatchTypeFuzzy,
+				},
+			},
+			expectedSQL:   "SELECT path, groupArray(DISTINCT type) AS types, max(last_seen) AS last_seen FROM signoz_metadata.distributed_field_keys WHERE (LOWER(path) LIKE LOWER(?) OR LOWER(path) LIKE LOWER(?)) GROUP BY path ORDER BY last_seen DESC LIMIT ?",
+			expectedArgs:  []any{"%user%", "%admin%", defaultPathLimit},
+			expectedLimit: defaultPathLimit,
+		},
+		{
+			name: "Search with Contains operator (should default to LIKE)",
+			fieldKeySelectors: []*telemetrytypes.FieldKeySelector{
+				{
+					Name:              "test",
+					SelectorMatchType: telemetrytypes.FieldSelectorMatchTypeFuzzy,
+				},
+			},
+			expectedSQL:   "SELECT path, groupArray(DISTINCT type) AS types, max(last_seen) AS last_seen FROM signoz_metadata.distributed_field_keys WHERE (LOWER(path) LIKE LOWER(?)) GROUP BY path ORDER BY last_seen DESC LIMIT ?",
+			expectedArgs:  []any{"%test%", defaultPathLimit},
+			expectedLimit: defaultPathLimit,
+		},
+	}
+
+	for _, tc := range testCases {
+		t.Run(tc.name, func(t *testing.T) {
+			query, args, limit := buildGetBodyJSONPathsQuery(tc.fieldKeySelectors)
+			require.Equal(t, tc.expectedSQL, query)
+			require.Equal(t, tc.expectedArgs, args)
+			require.Equal(t, tc.expectedLimit, limit)
+		})
+	}
+}
+
 func TestBuildListLogsJSONIndexesQuery(t *testing.T) {
 	testCases := []struct {
 		name         string

pkg/telemetrymetadata/metadata.go

@@ -368,20 +368,6 @@ func (t *telemetryMetaStore) logsTblStatementToFieldKeys(ctx context.Context) ([
 	return materialisedKeys, nil
 }
 
-// logKeysUnionArm declares one arm of the UNION ALL in getLogsKeys.
-// All per-table variance is captured here so the loop body can stay uniform.
-type logKeysUnionArm struct {
-	shouldQuery        bool
-	fieldContext       telemetrytypes.FieldContext
-	table              string
-	nameColumn         string                                    // column holding the field/key name (e.g., "name" or "field_name")
-	dataTypeColumn     string                                    // column used in WHERE/GROUP BY
-	dataTypeSelectExpr string                                    // expression used in SELECT (may wrap with lower())
-	addBaseFilters     func(sb *sqlbuilder.SelectBuilder)        // mandatory WHERE filters (e.g., signal, field_context)
-	encodeDataType     func(telemetrytypes.FieldDataType) string // how to render a FieldDataType in WHERE values
-	extraOrBranch      func(sb *sqlbuilder.SelectBuilder) string // optional extra OR branch (e.g., body parent-types)
-}
-
 // getLogsKeys returns the keys from the spans that match the field selection criteria.
 func (t *telemetryMetaStore) getLogsKeys(ctx context.Context, fieldKeySelectors []*telemetrytypes.FieldKeySelector) ([]*telemetrytypes.TelemetryFieldKey, bool, error) {
 	ctx = ctxtypes.NewContextWithCommentVals(ctx, map[string]string{
@@ -411,152 +397,90 @@ func (t *telemetryMetaStore) getLogsKeys(ctx context.Context, fieldKeySelectors
 	// tables to query based on field selectors
 	queryAttributeTable := false
 	queryResourceTable := false
-	queryBodyTable := false
 
 	for _, selector := range fieldKeySelectors {
 		if selector.FieldContext == telemetrytypes.FieldContextUnspecified {
-			// unspecified context, query all tables
+			// unspecified context, query both tables
 			queryAttributeTable = true
 			queryResourceTable = true
-			queryBodyTable = true
 			break
 		} else if selector.FieldContext == telemetrytypes.FieldContextAttribute {
 			queryAttributeTable = true
 		} else if selector.FieldContext == telemetrytypes.FieldContextResource {
 			queryResourceTable = true
-		} else if selector.FieldContext == telemetrytypes.FieldContextBody && querybuilder.BodyJSONQueryEnabled {
-			queryBodyTable = true
 		}
 	}
 
-	// body keys are gated behind the feature flag
-	queryBodyTable = queryBodyTable && querybuilder.BodyJSONQueryEnabled
-
-	// pre-compute parent array path names from body selectors for JSON plan building;
-	// these will be fetched as a separate UNION arm filtered to ArrayJSON/ArrayDynamic only
-	parentPaths := make(map[string]bool)
-	if queryBodyTable {
-		for _, sel := range fieldKeySelectors {
-			if sel.FieldContext != telemetrytypes.FieldContextBody &&
-				sel.FieldContext != telemetrytypes.FieldContextUnspecified {
-				continue
-			}
-			key := &telemetrytypes.TelemetryFieldKey{
-				Name:         sel.Name,
-				Signal:       telemetrytypes.SignalLogs,
-				FieldContext: telemetrytypes.FieldContextBody,
-			}
-			if !key.KeyNameContainsArray() {
-				continue
-			}
-			for _, parent := range key.ArrayParentPaths() {
-				parentPaths[parent] = true
-			}
-		}
+	tablesToQuery := []struct {
+		fieldContext telemetrytypes.FieldContext
+		shouldQuery  bool
+	}{
+		{telemetrytypes.FieldContextAttribute, queryAttributeTable},
+		{telemetrytypes.FieldContextResource, queryResourceTable},
 	}
 
-	// Each UNION arm differs only in: table, data-type column name and SELECT
-	// expression (lower-wrapped for historical mixed-case in attr/resource),
-	// base WHERE filters, the per-selector data-type encoding, and (for body)
-	// an extra OR branch that fetches parent array types for JSON plan building.
-	// All other logic is shared by the loop below.
-	tablesToQuery := []logKeysUnionArm{
-		{
-			shouldQuery:        queryAttributeTable,
-			fieldContext:       telemetrytypes.FieldContextAttribute,
-			table:              t.logsDBName + "." + t.logAttributeKeysTblName,
-			nameColumn:         "name",
-			dataTypeColumn:     "datatype",
-			dataTypeSelectExpr: "lower(datatype)",
-			addBaseFilters:     func(*sqlbuilder.SelectBuilder) {},
-			encodeDataType:     func(ft telemetrytypes.FieldDataType) string { return ft.TagDataType() },
-			extraOrBranch:      func(*sqlbuilder.SelectBuilder) string { return "" },
-		},
-		{
-			shouldQuery:        queryResourceTable,
-			fieldContext:       telemetrytypes.FieldContextResource,
-			table:              t.logsDBName + "." + t.logResourceKeysTblName,
-			nameColumn:         "name",
-			dataTypeColumn:     "datatype",
-			dataTypeSelectExpr: "lower(datatype)",
-			addBaseFilters:     func(*sqlbuilder.SelectBuilder) {},
-			encodeDataType:     func(ft telemetrytypes.FieldDataType) string { return ft.TagDataType() },
-			extraOrBranch:      func(*sqlbuilder.SelectBuilder) string { return "" },
-		},
-		{
-			shouldQuery:        queryBodyTable,
-			fieldContext:       telemetrytypes.FieldContextBody,
-			table:              fmt.Sprintf("%s.%s", DBName, FieldKeysTable),
-			nameColumn:         "field_name",
-			dataTypeColumn:     "field_data_type",
-			dataTypeSelectExpr: "field_data_type",
-			addBaseFilters: func(sb *sqlbuilder.SelectBuilder) {
-				sb.Where(sb.E("signal", telemetrytypes.SignalLogs.StringValue()))
-				sb.Where(sb.E("field_context", telemetrytypes.FieldContextBody.StringValue()))
-			},
-			encodeDataType: func(ft telemetrytypes.FieldDataType) string { return ft.StringValue() },
-			extraOrBranch: func(sb *sqlbuilder.SelectBuilder) string {
-				if len(parentPaths) == 0 {
-					return ""
-				}
-				names := make([]any, 0, len(parentPaths))
-				for n := range parentPaths {
-					names = append(names, n)
-				}
-				return sb.And(
-					sb.In("field_name", names...),
-					sb.In("field_data_type",
-						telemetrytypes.FieldDataTypeArrayDynamic.StringValue(),
-						telemetrytypes.FieldDataTypeArrayJSON.StringValue(),
-					),
-				)
-			},
-		},
-	}
-
-	for _, arm := range tablesToQuery {
-		if !arm.shouldQuery {
+	for _, table := range tablesToQuery {
+		if !table.shouldQuery {
 			continue
 		}
 
+		fieldContext := table.fieldContext
+
+		// table name based on field context
+		var tblName string
+		if fieldContext == telemetrytypes.FieldContextAttribute {
+			tblName = t.logsDBName + "." + t.logAttributeKeysTblName
+		} else {
+			tblName = t.logsDBName + "." + t.logResourceKeysTblName
+		}
+
 		sb := sqlbuilder.Select(
-			arm.nameColumn+" AS tag_key",
-			fmt.Sprintf("'%s' AS tag_type", arm.fieldContext.TagType()),
-			arm.dataTypeSelectExpr+" AS tag_data_type",
-			fmt.Sprintf("%d AS priority", getPriorityForContext(arm.fieldContext)),
-		).From(arm.table)
+			"name AS tag_key",
+			fmt.Sprintf("'%s' AS tag_type", fieldContext.TagType()),
+			"lower(datatype) AS tag_data_type", // in logs, we had some historical data with capital and small case
+			fmt.Sprintf(`%d AS priority`, getPriorityForContext(fieldContext)),
+		).From(tblName)
 
-		arm.addBaseFilters(sb)
+		var limit int
+		conds := []string{}
 
-		branches := []string{}
-		for _, sel := range fieldKeySelectors {
-			if sel.FieldContext != telemetrytypes.FieldContextUnspecified && sel.FieldContext != arm.fieldContext {
+		for _, fieldKeySelector := range fieldKeySelectors {
+			// Include this selector if:
+			// 1. It has unspecified context (matches all tables)
+			// 2. Its context matches the current table's context
+			if fieldKeySelector.FieldContext != telemetrytypes.FieldContextUnspecified &&
+				fieldKeySelector.FieldContext != fieldContext {
 				continue
 			}
-			parts := []string{}
-			if sel.SelectorMatchType == telemetrytypes.FieldSelectorMatchTypeExact {
-				parts = append(parts, sb.E(arm.nameColumn, sel.Name))
+
+			// key part of the selector
+			fieldKeyConds := []string{}
+			if fieldKeySelector.SelectorMatchType == telemetrytypes.FieldSelectorMatchTypeExact {
+				fieldKeyConds = append(fieldKeyConds, sb.E("name", fieldKeySelector.Name))
 			} else {
-				parts = append(parts, sb.ILike(arm.nameColumn, "%"+escapeForLike(sel.Name)+"%"))
+				fieldKeyConds = append(fieldKeyConds, sb.ILike("name", "%"+escapeForLike(fieldKeySelector.Name)+"%"))
 			}
-			if sel.FieldDataType != telemetrytypes.FieldDataTypeUnspecified {
-				parts = append(parts, sb.E(arm.dataTypeColumn, arm.encodeDataType(sel.FieldDataType)))
+
+			// now look at the field data type
+			if fieldKeySelector.FieldDataType != telemetrytypes.FieldDataTypeUnspecified {
+				fieldKeyConds = append(fieldKeyConds, sb.E("datatype", fieldKeySelector.FieldDataType.TagDataType()))
 			}
-			if len(parts) > 0 {
-				branches = append(branches, sb.And(parts...))
+
+			if len(fieldKeyConds) > 0 {
+				conds = append(conds, sb.And(fieldKeyConds...))
 			}
+			limit += fieldKeySelector.Limit
 		}
 
-		if extra := arm.extraOrBranch(sb); extra != "" {
-			branches = append(branches, extra)
+		if len(conds) > 0 {
+			sb.Where(sb.Or(conds...))
 		}
 
-		if len(branches) > 0 {
-			sb.Where(sb.Or(branches...))
+		sb.GroupBy("name", "datatype")
+		if limit == 0 {
+			limit = 1000
 		}
 
-		sb.GroupBy(arm.nameColumn, arm.dataTypeColumn)
-
 		query, args := sb.BuildWithFlavor(sqlbuilder.ClickHouse)
 		queries = append(queries, query)
 		allArgs = append(allArgs, args...)
@@ -593,7 +517,6 @@ func (t *telemetryMetaStore) getLogsKeys(ctx context.Context, fieldKeySelectors
 	defer rows.Close()
 
 	keys := []*telemetrytypes.TelemetryFieldKey{}
-	parentTypeCache := make(map[string][]telemetrytypes.FieldDataType)
 	rowCount := 0
 	searchTexts := []string{}
 
@@ -617,17 +540,6 @@ func (t *telemetryMetaStore) getLogsKeys(ctx context.Context, fieldKeySelectors
 		if err != nil {
 			return nil, false, errors.Wrap(err, errors.TypeInternal, errors.CodeInternal, ErrFailedToGetLogsKeys.Error())
 		}
-
-		// body keys with ArrayJSON/ArrayDynamic types are internal container types
-		// used only for JSON access plan building; route to parentTypeCache, not to results
-		switch fieldDataType {
-		case telemetrytypes.FieldDataTypeArrayJSON, telemetrytypes.FieldDataTypeArrayDynamic:
-			if fieldContext == telemetrytypes.FieldContextBody && parentPaths[name] {
-				parentTypeCache[name] = append(parentTypeCache[name], fieldDataType)
-				continue
-			}
-		}
-
 		key, ok := mapOfKeys[name+";"+fieldContext.StringValue()+";"+fieldDataType.StringValue()]
 
 		// if there is no materialised column, create a key with the field context and data type
@@ -681,11 +593,13 @@ func (t *telemetryMetaStore) getLogsKeys(ctx context.Context, fieldKeySelectors
 		}
 	}
 
-	// enrich body keys with promoted paths, indexes, and JSON access plans
 	if querybuilder.BodyJSONQueryEnabled {
-		if err := t.enrichJSONKeys(ctx, fieldKeySelectors, keys, parentTypeCache); err != nil {
-			return nil, false, err
+		bodyJSONPaths, finished, err := t.buildBodyJSONPaths(ctx, fieldKeySelectors) // LIKE for pattern matching
+		if err != nil {
+			t.logger.ErrorContext(ctx, "failed to extract body JSON paths", errors.Attr(err))
 		}
+		keys = append(keys, bodyJSONPaths...)
+		complete = complete && finished
 	}
 
 	if _, err := t.updateColumnEvolutionMetadataForKeys(ctx, keys); err != nil {

pkg/telemetrymetadata/tables.go

@@ -1,13 +1,13 @@
 package telemetrymetadata
 
-import otelconst "github.com/SigNoz/signoz-otel-collector/constants"
+import otelcollectorconst "github.com/SigNoz/signoz-otel-collector/constants"
 
 const (
 	DBName                           = "signoz_metadata"
 	AttributesMetadataTableName      = "distributed_attributes_metadata"
 	AttributesMetadataLocalTableName = "attributes_metadata"
 	ColumnEvolutionMetadataTableName = "distributed_column_evolution_metadata"
-	FieldKeysTable                   = otelconst.DistributedFieldKeysTable
+	PathTypesTableName               = otelcollectorconst.DistributedFieldKeysTable
 	// Column Evolution table stores promoted paths as (signal, column_name, field_context, field_name); see signoz-otel-collector metadata_migrations.
 	PromotedPathsTableName = "distributed_column_evolution_metadata"
 	SkipIndexTableName     = "system.data_skipping_indices"

pkg/types/telemetrytypes/field.go

@@ -37,6 +37,7 @@ type TelemetryFieldKey struct {
 	FieldContext  FieldContext  `json:"fieldContext,omitzero"`
 	FieldDataType FieldDataType `json:"fieldDataType,omitzero"`
 
+	JSONDataType *JSONDataType       `json:"-"`
 	JSONPlan     JSONAccessPlan      `json:"-"`
 	Indexes      []JSONDataTypeIndex `json:"-"`
 	Materialized bool                `json:"-"` // refers to promoted in case of body.... fields
@@ -79,11 +80,6 @@ func (f *TelemetryFieldKey) ArrayParentSelectors() []*FieldKeySelector {
 	return selectors
 }
 
-// GetJSONDataType derives the JSONDataType from FieldDataType.
-func (f *TelemetryFieldKey) GetJSONDataType() JSONDataType {
-	return MappingFieldDataTypeToJSONDataType[f.FieldDataType]
-}
-
 func (f TelemetryFieldKey) String() string {
 	var sb strings.Builder
 	fmt.Fprintf(&sb, "name=%s", f.Name)
@@ -96,6 +92,9 @@ func (f TelemetryFieldKey) String() string {
 	if f.Materialized {
 		sb.WriteString(",materialized=true")
 	}
+	if f.JSONDataType != nil {
+		fmt.Fprintf(&sb, ",jsondatatype=%s", f.JSONDataType.StringValue())
+	}
 	if len(f.Indexes) > 0 {
 		sb.WriteString(",indexes=[")
 		for i, index := range f.Indexes {
@@ -118,6 +117,7 @@ func (f TelemetryFieldKey) Text() string {
 func (f *TelemetryFieldKey) OverrideMetadataFrom(src *TelemetryFieldKey) {
 	f.FieldContext = src.FieldContext
 	f.FieldDataType = src.FieldDataType
+	f.JSONDataType = src.JSONDataType
 	f.Indexes = src.Indexes
 	f.Materialized = src.Materialized
 	f.JSONPlan = src.JSONPlan

pkg/types/telemetrytypes/field_datatype.go

@@ -31,7 +31,7 @@ var (
 	FieldDataTypeArrayInt64  = FieldDataType{valuer.NewString("[]int64")}
 	FieldDataTypeArrayNumber = FieldDataType{valuer.NewString("[]number")}
 
-	FieldDataTypeArrayJSON    = FieldDataType{valuer.NewString("[]json")}
+	FieldDataTypeArrayObject  = FieldDataType{valuer.NewString("[]object")}
 	FieldDataTypeArrayDynamic = FieldDataType{valuer.NewString("[]dynamic")}
 
 	// Map string representations to FieldDataType values
@@ -72,8 +72,6 @@ var (
 		"[]float64": FieldDataTypeArrayFloat64,
 		"[]number":  FieldDataTypeArrayNumber,
 		"[]bool":    FieldDataTypeArrayBool,
-		"[]json":    FieldDataTypeArrayJSON,
-		"[]dynamic": FieldDataTypeArrayDynamic,
 
 		// c-style array types
 		"string[]":  FieldDataTypeArrayString,
@@ -81,8 +79,6 @@ var (
 		"float64[]": FieldDataTypeArrayFloat64,
 		"number[]":  FieldDataTypeArrayNumber,
 		"bool[]":    FieldDataTypeArrayBool,
-		"json[]":    FieldDataTypeArrayJSON,
-		"dynamic[]": FieldDataTypeArrayDynamic,
 	}
 
 	fieldDataTypeToCHDataType = map[FieldDataType]string{

pkg/types/telemetrytypes/json_access_plan.go

@@ -19,8 +19,7 @@ var (
 	BranchJSON    = JSONAccessBranchType{valuer.NewString("json")}
 	BranchDynamic = JSONAccessBranchType{valuer.NewString("dynamic")}
 
-	CodePlanIndexOutOfBounds     = errors.MustNewCode("plan_index_out_of_bounds")
-	CodePlanFieldDataTypeMissing = errors.MustNewCode("field_data_type_missing")
+	CodePlanIndexOutOfBounds = errors.MustNewCode("plan_index_out_of_bounds")
 )
 
 type JSONColumnMetadata struct {
@@ -43,6 +42,9 @@ type JSONAccessNode struct {
 	IsTerminal bool
 	isRoot     bool // marked true for only body_v2 and body_promoted
 
+	// Precomputed type information (single source of truth)
+	AvailableTypes []JSONDataType
+
 	// Array type branches (Array(JSON) vs Array(Dynamic))
 	Branches map[JSONAccessBranchType]*JSONAccessNode
 
@@ -104,7 +106,7 @@ type planBuilder struct {
 	paths      []string // cumulative paths for type cache lookups
 	segments   []string // individual path segments for node names
 	isPromoted bool
-	typeCache  map[string][]FieldDataType
+	typeCache  map[string][]JSONDataType
 }
 
 // buildPlan recursively builds the path plan tree.
@@ -136,41 +138,34 @@ func (pb *planBuilder) buildPlan(index int, parent *JSONAccessNode, isDynArrChil
 		}
 	}
 
+	// Use cached types from the batched metadata query
+	types, ok := pb.typeCache[pathSoFar]
+	if !ok {
+		return nil, errors.NewInternalf(errors.CodeInvalidInput, "types missing for path %s", pathSoFar)
+	}
+
 	// Create node for this path segment
 	node := &JSONAccessNode{
 		Name:            segmentName,
 		IsTerminal:      isTerminal,
+		AvailableTypes:  types,
 		Branches:        make(map[JSONAccessBranchType]*JSONAccessNode),
 		Parent:          parent,
 		MaxDynamicTypes: maxTypes,
 		MaxDynamicPaths: maxPaths,
 	}
 
+	hasJSON := slices.Contains(node.AvailableTypes, ArrayJSON)
+	hasDynamic := slices.Contains(node.AvailableTypes, ArrayDynamic)
+
 	// Configure terminal if this is the last part
 	if isTerminal {
-		// fielddatatype must not be unspecified else expression can not be generated
-		if pb.key.FieldDataType == FieldDataTypeUnspecified {
-			return nil, errors.NewInternalf(CodePlanFieldDataTypeMissing, "field data type is missing for path %s", pathSoFar)
-		}
-
 		node.TerminalConfig = &TerminalConfig{
 			Key:      pb.key,
-			ElemType: pb.key.GetJSONDataType(),
+			ElemType: *pb.key.JSONDataType,
 		}
 	} else {
 		var err error
-		// Use cached types from the batched metadata query
-		types, ok := pb.typeCache[pathSoFar]
-		if !ok {
-			return nil, errors.NewInternalf(errors.CodeInvalidInput, "types missing for path %s", pathSoFar)
-		}
-
-		hasJSON := slices.Contains(types, FieldDataTypeArrayJSON)
-		hasDynamic := slices.Contains(types, FieldDataTypeArrayDynamic)
-		if !hasJSON && !hasDynamic {
-			return nil, errors.NewInternalf(CodePlanFieldDataTypeMissing, "array data type missing for path %s", pathSoFar)
-		}
-
 		if hasJSON {
 			node.Branches[BranchJSON], err = pb.buildPlan(index+1, node, false)
 			if err != nil {
@@ -190,7 +185,7 @@ func (pb *planBuilder) buildPlan(index int, parent *JSONAccessNode, isDynArrChil
 
 // buildJSONAccessPlan builds a tree structure representing the complete JSON path traversal
 // that precomputes all possible branches and their types.
-func (key *TelemetryFieldKey) SetJSONAccessPlan(columnInfo JSONColumnMetadata, typeCache map[string][]FieldDataType,
+func (key *TelemetryFieldKey) SetJSONAccessPlan(columnInfo JSONColumnMetadata, typeCache map[string][]JSONDataType,
 ) error {
 	// if path is empty, return nil
 	if key.Name == "" {

pkg/types/telemetrytypes/json_access_plan_test.go

@@ -19,11 +19,11 @@ const (
 // ============================================================================
 
 // makeKey creates a TelemetryFieldKey for testing.
-func makeKey(name string, dataType FieldDataType, materialized bool) *TelemetryFieldKey {
+func makeKey(name string, dataType JSONDataType, materialized bool) *TelemetryFieldKey {
 	return &TelemetryFieldKey{
-		Name:          name,
-		FieldDataType: dataType,
-		Materialized:  materialized,
+		Name:         name,
+		JSONDataType: &dataType,
+		Materialized: materialized,
 	}
 }
 
@@ -65,6 +65,14 @@ func toTestNode(n *JSONAccessNode) *jsonAccessTestNode {
 		out.Column = n.Parent.Name
 	}
 
+	// AvailableTypes as strings (using StringValue for stable representation)
+	if len(n.AvailableTypes) > 0 {
+		out.AvailableTypes = make([]string, 0, len(n.AvailableTypes))
+		for _, t := range n.AvailableTypes {
+			out.AvailableTypes = append(out.AvailableTypes, t.StringValue())
+		}
+	}
+
 	// Terminal config
 	if n.TerminalConfig != nil {
 		out.ElemType = n.TerminalConfig.ElemType.StringValue()
@@ -234,12 +242,12 @@ func TestPlanJSON_BasicStructure(t *testing.T) {
 	}{
 		{
 			name: "Simple path not promoted",
-			key:  makeKey("user.name", FieldDataTypeString, false),
+			key:  makeKey("user.name", String, false),
 			expectedYAML: fmt.Sprintf(`
 - name: user.name
   column: %s
   availableTypes:
-    - string
+    - String
   maxDynamicTypes: 16
   isTerminal: true
   elemType: String
@@ -247,19 +255,19 @@ func TestPlanJSON_BasicStructure(t *testing.T) {
 		},
 		{
 			name: "Simple path promoted",
-			key:  makeKey("user.name", FieldDataTypeString, true),
+			key:  makeKey("user.name", String, true),
 			expectedYAML: fmt.Sprintf(`
 - name: user.name
   column: %s
   availableTypes:
-    - string
+    - String
   maxDynamicTypes: 16
   isTerminal: true
   elemType: String
 - name: user.name
   column: %s
   availableTypes:
-    - string
+    - String
   maxDynamicTypes: 16
   maxDynamicPaths: 256
   isTerminal: true
@@ -268,7 +276,7 @@ func TestPlanJSON_BasicStructure(t *testing.T) {
 		},
 		{
 			name:         "Empty path returns error",
-			key:          makeKey("", FieldDataTypeString, false),
+			key:          makeKey("", String, false),
 			expectErr:    true,
 			expectedYAML: "",
 		},
@@ -307,13 +315,13 @@ func TestPlanJSON_ArrayPaths(t *testing.T) {
 - name: education
   column: %s
   availableTypes:
-    - "[]json"
+    - Array(JSON)
   maxDynamicTypes: 16
   branches:
     json:
       name: name
       availableTypes:
-        - string
+        - String
       maxDynamicTypes: 8
       isTerminal: true
       elemType: String
@@ -326,27 +334,27 @@ func TestPlanJSON_ArrayPaths(t *testing.T) {
 - name: education
   column: %s
   availableTypes:
-    - "[]json"
+    - Array(JSON)
   maxDynamicTypes: 16
   branches:
     json:
       name: awards
       availableTypes:
-        - "[]dynamic"
-        - "[]json"
+        - Array(Dynamic)
+        - Array(JSON)
       maxDynamicTypes: 8
       branches:
         json:
           name: type
           availableTypes:
-            - string
+            - String
           maxDynamicTypes: 4
           isTerminal: true
           elemType: String
         dynamic:
           name: type
           availableTypes:
-            - string
+            - String
           maxDynamicTypes: 16
           maxDynamicPaths: 256
           isTerminal: true
@@ -360,42 +368,42 @@ func TestPlanJSON_ArrayPaths(t *testing.T) {
 - name: interests
   column: %s
   availableTypes:
-    - "[]json"
+    - Array(JSON)
   maxDynamicTypes: 16
   branches:
     json:
       name: entities
       availableTypes:
-        - "[]json"
+        - Array(JSON)
       maxDynamicTypes: 8
       branches:
         json:
           name: reviews
           availableTypes:
-            - "[]json"
+            - Array(JSON)
           maxDynamicTypes: 4
           branches:
             json:
               name: entries
               availableTypes:
-                - "[]json"
+                - Array(JSON)
               maxDynamicTypes: 2
               branches:
                 json:
                   name: metadata
                   availableTypes:
-                    - "[]json"
+                    - Array(JSON)
                   maxDynamicTypes: 1
                   branches:
                     json:
                       name: positions
                       availableTypes:
-                        - "[]json"
+                        - Array(JSON)
                       branches:
                         json:
                           name: name
                           availableTypes:
-                            - string
+                            - String
                           isTerminal: true
                           elemType: String
 `, bodyV2Column),
@@ -407,13 +415,13 @@ func TestPlanJSON_ArrayPaths(t *testing.T) {
 - name: education
   column: %s
   availableTypes:
-    - "[]json"
+    - Array(JSON)
   maxDynamicTypes: 16
   branches:
     json:
       name: name
       availableTypes:
-        - string
+        - String
       maxDynamicTypes: 8
       isTerminal: true
       elemType: String
@@ -423,7 +431,7 @@ func TestPlanJSON_ArrayPaths(t *testing.T) {
 
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
-			key := makeKey(tt.path, FieldDataTypeString, false)
+			key := makeKey(tt.path, String, false)
 			err := key.SetJSONAccessPlan(JSONColumnMetadata{
 				BaseColumn:     bodyV2Column,
 				PromotedColumn: bodyPromotedColumn,
@@ -442,7 +450,7 @@ func TestPlanJSON_PromotedVsNonPromoted(t *testing.T) {
 	path := "education[].awards[].type"
 
 	t.Run("Non-promoted plan", func(t *testing.T) {
-		key := makeKey(path, FieldDataTypeString, false)
+		key := makeKey(path, String, false)
 		err := key.SetJSONAccessPlan(JSONColumnMetadata{
 			BaseColumn:     bodyV2Column,
 			PromotedColumn: bodyPromotedColumn,
@@ -454,27 +462,27 @@ func TestPlanJSON_PromotedVsNonPromoted(t *testing.T) {
 - name: education
   column: %s
   availableTypes:
-    - "[]json"
+    - Array(JSON)
   maxDynamicTypes: 16
   branches:
     json:
       name: awards
       availableTypes:
-        - "[]dynamic"
-        - "[]json"
+        - Array(Dynamic)
+        - Array(JSON)
       maxDynamicTypes: 8
       branches:
         json:
           name: type
           availableTypes:
-            - string
+            - String
           maxDynamicTypes: 4
           isTerminal: true
           elemType: String
         dynamic:
           name: type
           availableTypes:
-            - string
+            - String
           maxDynamicTypes: 16
           maxDynamicPaths: 256
           isTerminal: true
@@ -485,7 +493,7 @@ func TestPlanJSON_PromotedVsNonPromoted(t *testing.T) {
 	})
 
 	t.Run("Promoted plan", func(t *testing.T) {
-		key := makeKey(path, FieldDataTypeString, true)
+		key := makeKey(path, String, true)
 		err := key.SetJSONAccessPlan(JSONColumnMetadata{
 			BaseColumn:     bodyV2Column,
 			PromotedColumn: bodyPromotedColumn,
@@ -497,27 +505,27 @@ func TestPlanJSON_PromotedVsNonPromoted(t *testing.T) {
 - name: education
   column: %s
   availableTypes:
-    - "[]json"
+    - Array(JSON)
   maxDynamicTypes: 16
   branches:
     json:
       name: awards
       availableTypes:
-        - "[]dynamic"
-        - "[]json"
+        - Array(Dynamic)
+        - Array(JSON)
       maxDynamicTypes: 8
       branches:
         json:
           name: type
           availableTypes:
-            - string
+            - String
           maxDynamicTypes: 4
           isTerminal: true
           elemType: String
         dynamic:
           name: type
           availableTypes:
-            - string
+            - String
           maxDynamicTypes: 16
           maxDynamicPaths: 256
           isTerminal: true
@@ -525,22 +533,22 @@ func TestPlanJSON_PromotedVsNonPromoted(t *testing.T) {
 - name: education
   column: %s
   availableTypes:
-    - "[]json"
+    - Array(JSON)
   maxDynamicTypes: 16
   maxDynamicPaths: 256
   branches:
     json:
       name: awards
       availableTypes:
-        - "[]dynamic"
-        - "[]json"
+        - Array(Dynamic)
+        - Array(JSON)
       maxDynamicTypes: 8
       maxDynamicPaths: 64
       branches:
         json:
           name: type
           availableTypes:
-            - string
+            - String
           maxDynamicTypes: 4
           maxDynamicPaths: 16
           isTerminal: true
@@ -548,7 +556,7 @@ func TestPlanJSON_PromotedVsNonPromoted(t *testing.T) {
         dynamic:
           name: type
           availableTypes:
-            - string
+            - String
           maxDynamicTypes: 16
           maxDynamicPaths: 256
           isTerminal: true
@@ -580,42 +588,42 @@ func TestPlanJSON_EdgeCases(t *testing.T) {
 - name: interests
   column: %s
   availableTypes:
-    - "[]json"
+    - Array(JSON)
   maxDynamicTypes: 16
   branches:
     json:
       name: entities
       availableTypes:
-        - "[]json"
+        - Array(JSON)
       maxDynamicTypes: 8
       branches:
         json:
           name: reviews
           availableTypes:
-            - "[]json"
+            - Array(JSON)
           maxDynamicTypes: 4
           branches:
             json:
               name: entries
               availableTypes:
-                - "[]json"
+                - Array(JSON)
               maxDynamicTypes: 2
               branches:
                 json:
                   name: metadata
                   availableTypes:
-                    - "[]json"
+                    - Array(JSON)
                   maxDynamicTypes: 1
                   branches:
                     json:
                       name: positions
                       availableTypes:
-                        - "[]json"
+                        - Array(JSON)
                       branches:
                         json:
                           name: name
                           availableTypes:
-                            - string
+                            - String
                           isTerminal: true
                           elemType: String
 `, bodyV2Column),
@@ -627,14 +635,14 @@ func TestPlanJSON_EdgeCases(t *testing.T) {
 - name: education
   column: %s
   availableTypes:
-    - "[]json"
+    - Array(JSON)
   maxDynamicTypes: 16
   branches:
     json:
       name: type
       availableTypes:
-        - string
-        - int64
+        - String
+        - Int64
       maxDynamicTypes: 8
       isTerminal: true
       elemType: String
@@ -647,7 +655,7 @@ func TestPlanJSON_EdgeCases(t *testing.T) {
 - name: education
   column: %s
   availableTypes:
-    - "[]json"
+    - Array(JSON)
   maxDynamicTypes: 16
   isTerminal: true
   elemType: Array(JSON)
@@ -658,12 +666,12 @@ func TestPlanJSON_EdgeCases(t *testing.T) {
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
 			// Choose key type based on path; operator does not affect the tree shape asserted here.
-			keyType := FieldDataTypeString
+			keyType := String
 			switch tt.path {
 			case "education":
-				keyType = FieldDataTypeArrayJSON
+				keyType = ArrayJSON
 			case "education[].type":
-				keyType = FieldDataTypeString
+				keyType = String
 			}
 			key := makeKey(tt.path, keyType, false)
 			err := key.SetJSONAccessPlan(JSONColumnMetadata{
@@ -684,7 +692,7 @@ func TestPlanJSON_EdgeCases(t *testing.T) {
 func TestPlanJSON_TreeStructure(t *testing.T) {
 	types, _ := TestJSONTypeSet()
 	path := "education[].awards[].participated[].team[].branch"
-	key := makeKey(path, FieldDataTypeString, false)
+	key := makeKey(path, String, false)
 	err := key.SetJSONAccessPlan(JSONColumnMetadata{
 		BaseColumn:     bodyV2Column,
 		PromotedColumn: bodyPromotedColumn,
@@ -696,47 +704,47 @@ func TestPlanJSON_TreeStructure(t *testing.T) {
 - name: education
   column: %s
   availableTypes:
-    - "[]json"
+    - Array(JSON)
   maxDynamicTypes: 16
   branches:
     json:
       name: awards
       availableTypes:
-        - "[]dynamic"
-        - "[]json"
+        - Array(Dynamic)
+        - Array(JSON)
       maxDynamicTypes: 8
       branches:
         json:
           name: participated
           availableTypes:
-            - "[]dynamic"
-            - "[]json"
+            - Array(Dynamic)
+            - Array(JSON)
           maxDynamicTypes: 4
           branches:
             json:
               name: team
               availableTypes:
-                - "[]json"
+                - Array(JSON)
               maxDynamicTypes: 2
               branches:
                 json:
                   name: branch
                   availableTypes:
-                    - string
+                    - String
                   maxDynamicTypes: 1
                   isTerminal: true
                   elemType: String
             dynamic:
               name: team
               availableTypes:
-                - "[]json"
+                - Array(JSON)
               maxDynamicTypes: 16
               maxDynamicPaths: 256
               branches:
                 json:
                   name: branch
                   availableTypes:
-                    - string
+                    - String
                   maxDynamicTypes: 8
                   maxDynamicPaths: 64
                   isTerminal: true
@@ -744,22 +752,22 @@ func TestPlanJSON_TreeStructure(t *testing.T) {
         dynamic:
           name: participated
           availableTypes:
-            - "[]dynamic"
-            - "[]json"
+            - Array(Dynamic)
+            - Array(JSON)
           maxDynamicTypes: 16
           maxDynamicPaths: 256
           branches:
             json:
               name: team
               availableTypes:
-                - "[]json"
+                - Array(JSON)
               maxDynamicTypes: 8
               maxDynamicPaths: 64
               branches:
                 json:
                   name: branch
                   availableTypes:
-                    - string
+                    - String
                   maxDynamicTypes: 4
                   maxDynamicPaths: 16
                   isTerminal: true
@@ -767,14 +775,14 @@ func TestPlanJSON_TreeStructure(t *testing.T) {
             dynamic:
               name: team
               availableTypes:
-                - "[]json"
+                - Array(JSON)
               maxDynamicTypes: 16
               maxDynamicPaths: 256
               branches:
                 json:
                   name: branch
                   availableTypes:
-                    - string
+                    - String
                   maxDynamicTypes: 8
                   maxDynamicPaths: 64
                   isTerminal: true

pkg/types/telemetrytypes/json_datatype.go

@@ -46,6 +46,14 @@ var MappingStringToJSONDataType = map[string]JSONDataType{
 	"Array(JSON)":              ArrayJSON,
 }
 
+var ScalerTypeToArrayType = map[JSONDataType]JSONDataType{
+	String:  ArrayString,
+	Int64:   ArrayInt64,
+	Float64: ArrayFloat64,
+	Bool:    ArrayBool,
+	Dynamic: ArrayDynamic,
+}
+
 var MappingFieldDataTypeToJSONDataType = map[FieldDataType]JSONDataType{
 	FieldDataTypeString:       String,
 	FieldDataTypeInt64:        Int64,
@@ -55,8 +63,18 @@ var MappingFieldDataTypeToJSONDataType = map[FieldDataType]JSONDataType{
 	FieldDataTypeArrayString:  ArrayString,
 	FieldDataTypeArrayInt64:   ArrayInt64,
 	FieldDataTypeArrayFloat64: ArrayFloat64,
-	FieldDataTypeArrayNumber:  ArrayFloat64,
 	FieldDataTypeArrayBool:    ArrayBool,
-	FieldDataTypeArrayDynamic: ArrayDynamic,
-	FieldDataTypeArrayJSON:    ArrayJSON,
+}
+
+var MappingJSONDataTypeToFieldDataType = map[JSONDataType]FieldDataType{
+	String:       FieldDataTypeString,
+	Int64:        FieldDataTypeInt64,
+	Float64:      FieldDataTypeFloat64,
+	Bool:         FieldDataTypeBool,
+	ArrayString:  FieldDataTypeArrayString,
+	ArrayInt64:   FieldDataTypeArrayInt64,
+	ArrayFloat64: FieldDataTypeArrayFloat64,
+	ArrayBool:    FieldDataTypeArrayBool,
+	ArrayDynamic: FieldDataTypeArrayDynamic,
+	ArrayJSON:    FieldDataTypeArrayObject,
 }

pkg/types/telemetrytypes/test_data.go

@@ -4,69 +4,69 @@ package telemetrytypes
 // Test JSON Type Set Data Setup
 // ============================================================================
 
-// TestJSONTypeSet returns a map of path->field data types for testing.
+// TestJSONTypeSet returns a map of path->types for testing.
 // This represents the type information available in the test JSON structure.
-func TestJSONTypeSet() (map[string][]FieldDataType, MetadataStore) {
-	types := map[string][]FieldDataType{
+func TestJSONTypeSet() (map[string][]JSONDataType, MetadataStore) {
+	types := map[string][]JSONDataType{
 
 		// ── user (primitives) ─────────────────────────────────────────────
-		"user.name":        {FieldDataTypeString},
-		"user.permissions": {FieldDataTypeArrayString},
-		"user.age":         {FieldDataTypeInt64, FieldDataTypeString}, // Int64/String ambiguity
-		"user.height":      {FieldDataTypeFloat64},
-		"user.active":      {FieldDataTypeBool}, // Bool — not IndexSupported
+		"user.name":        {String},
+		"user.permissions": {ArrayString},
+		"user.age":         {Int64, String}, // Int64/String ambiguity
+		"user.height":      {Float64},
+		"user.active":      {Bool}, // Bool — not IndexSupported
 
 		// Deeper non-array nesting (a.b.c — no array hops)
-		"user.address.zip": {FieldDataTypeInt64},
+		"user.address.zip": {Int64},
 
 		// ── education[] ───────────────────────────────────────────────────
 		// Pattern: x[].y
-		"education":              {FieldDataTypeArrayJSON},
-		"education[].name":       {FieldDataTypeString},
-		"education[].type":       {FieldDataTypeString, FieldDataTypeInt64},
-		"education[].year":       {FieldDataTypeInt64},
-		"education[].scores":     {FieldDataTypeArrayInt64},
-		"education[].parameters": {FieldDataTypeArrayFloat64, FieldDataTypeArrayDynamic},
+		"education":              {ArrayJSON},
+		"education[].name":       {String},
+		"education[].type":       {String, Int64},
+		"education[].year":       {Int64},
+		"education[].scores":     {ArrayInt64},
+		"education[].parameters": {ArrayFloat64, ArrayDynamic},
 
 		// Pattern: x[].y[]
-		"education[].awards": {FieldDataTypeArrayDynamic, FieldDataTypeArrayJSON},
+		"education[].awards": {ArrayDynamic, ArrayJSON},
 
 		// Pattern: x[].y[].z
-		"education[].awards[].name":     {FieldDataTypeString},
-		"education[].awards[].type":     {FieldDataTypeString},
-		"education[].awards[].semester": {FieldDataTypeInt64},
+		"education[].awards[].name":     {String},
+		"education[].awards[].type":     {String},
+		"education[].awards[].semester": {Int64},
 
 		// Pattern: x[].y[].z[]
-		"education[].awards[].participated": {FieldDataTypeArrayDynamic, FieldDataTypeArrayJSON},
+		"education[].awards[].participated": {ArrayDynamic, ArrayJSON},
 
 		// Pattern: x[].y[].z[].w
-		"education[].awards[].participated[].members": {FieldDataTypeArrayString},
+		"education[].awards[].participated[].members": {ArrayString},
 
 		// Pattern: x[].y[].z[].w[]
-		"education[].awards[].participated[].team": {FieldDataTypeArrayJSON},
+		"education[].awards[].participated[].team": {ArrayJSON},
 
 		// Pattern: x[].y[].z[].w[].v
-		"education[].awards[].participated[].team[].branch": {FieldDataTypeString},
+		"education[].awards[].participated[].team[].branch": {String},
 
 		// ── interests[] ───────────────────────────────────────────────────
-		"interests":                                                                 {FieldDataTypeArrayJSON},
-		"interests[].entities":                                                      {FieldDataTypeArrayJSON},
-		"interests[].entities[].reviews":                                            {FieldDataTypeArrayJSON},
-		"interests[].entities[].reviews[].entries":                                  {FieldDataTypeArrayJSON},
-		"interests[].entities[].reviews[].entries[].metadata":                       {FieldDataTypeArrayJSON},
-		"interests[].entities[].reviews[].entries[].metadata[].positions":           {FieldDataTypeArrayJSON},
-		"interests[].entities[].reviews[].entries[].metadata[].positions[].name":    {FieldDataTypeString},
-		"interests[].entities[].reviews[].entries[].metadata[].positions[].ratings": {FieldDataTypeArrayInt64, FieldDataTypeArrayString},
-		"http-events":                     {FieldDataTypeArrayJSON},
-		"http-events[].request-info.host": {FieldDataTypeString},
-		"ids":                             {FieldDataTypeArrayDynamic},
+		"interests":                                                                 {ArrayJSON},
+		"interests[].entities":                                                      {ArrayJSON},
+		"interests[].entities[].reviews":                                            {ArrayJSON},
+		"interests[].entities[].reviews[].entries":                                  {ArrayJSON},
+		"interests[].entities[].reviews[].entries[].metadata":                       {ArrayJSON},
+		"interests[].entities[].reviews[].entries[].metadata[].positions":           {ArrayJSON},
+		"interests[].entities[].reviews[].entries[].metadata[].positions[].name":    {String},
+		"interests[].entities[].reviews[].entries[].metadata[].positions[].ratings": {ArrayInt64, ArrayString},
+		"http-events":                     {ArrayJSON},
+		"http-events[].request-info.host": {String},
+		"ids":                             {ArrayDynamic},
 
 		// ── top-level primitives ──────────────────────────────────────────
-		"message":     {FieldDataTypeString},
-		"http-status": {FieldDataTypeInt64, FieldDataTypeString}, // hyphen in root key, ambiguous
+		"message":     {String},
+		"http-status": {Int64, String}, // hyphen in root key, ambiguous
 
 		// ── top-level nested objects (no array hops) ───────────────────────
-		"response.time-taken": {FieldDataTypeFloat64}, // hyphen inside nested key
+		"response.time-taken": {Float64}, // hyphen inside nested key
 	}
 
 	return types, nil