fix: linting issues

Co-authored-by: Tushar Vats <tushar@signoz.io>

pkg/instrumentation/sdk.go

@@ -14,7 +14,7 @@ import (
 	sdkmetric "go.opentelemetry.io/otel/metric"
 	sdkmetricnoop "go.opentelemetry.io/otel/metric/noop"
 	sdkresource "go.opentelemetry.io/otel/sdk/resource"
-	semconv "go.opentelemetry.io/otel/semconv/v1.40.0"
+	semconv "go.opentelemetry.io/otel/semconv/v1.39.0"
 	sdktrace "go.opentelemetry.io/otel/trace"
 )
 

pkg/querier/consume.go

@@ -419,6 +419,7 @@ func readAsRaw(rows driver.Rows, queryName string) (*qbtypes.RawData, error) {
 
 			rr.Data[name] = val
 		}
+		mergeSpanAttributeColumns(rr.Data)
 		outRows = append(outRows, &rr)
 	}
 	if err := rows.Err(); err != nil {
@@ -431,6 +432,48 @@ func readAsRaw(rows driver.Rows, queryName string) (*qbtypes.RawData, error) {
 	}, nil
 }
 
+// mergeSpanAttributeColumns merges the typed ClickHouse span attribute columns
+// (attributes_string, attributes_number, attributes_bool, resources_string) into
+// unified "attributes" and "resource_attributes" keys, removing the raw columns.
+// It is a no-op if none of the raw columns are present.
+func mergeSpanAttributeColumns(data map[string]any) {
+	attrStr, hasStr := data["attributes_string"]
+	attrNum, hasNum := data["attributes_number"]
+	attrBool, hasBool := data["attributes_bool"]
+	// todo(nitya): move to resource json
+	resStr, hasRes := data["resources_string"]
+
+	if !hasStr && !hasNum && !hasBool && !hasRes {
+		return
+	}
+
+	attributes := make(map[string]any)
+	if m, ok := attrStr.(map[string]string); ok {
+		for k, v := range m {
+			attributes[k] = v
+		}
+	}
+	if m, ok := attrNum.(map[string]float64); ok {
+		for k, v := range m {
+			attributes[k] = v
+		}
+	}
+	if m, ok := attrBool.(map[string]bool); ok {
+		for k, v := range m {
+			attributes[k] = v
+		}
+	}
+	delete(data, "attributes_string")
+	delete(data, "attributes_number")
+	delete(data, "attributes_bool")
+	data["attributes"] = attributes
+
+	if m, ok := resStr.(map[string]string); ok {
+		data["resource"] = m
+	}
+	delete(data, "resources_string")
+}
+
 // numericAsFloat converts numeric types to float64 efficiently.
 func numericAsFloat(v any) float64 {
 	switch x := v.(type) {

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_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))]."},
 			},
 		},
 		{
@@ -946,16 +946,16 @@ func buildTestTelemetryMetadataStore(t *testing.T, addIndexes bool) *telemetryty
 	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

@@ -21,72 +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")
 )
 
-// enrichBodyKeys 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.
-func (t *telemetryMetaStore) enrichBodyKeys(ctx context.Context, keys []*telemetrytypes.TelemetryFieldKey, parentTypeCache map[string][]telemetrytypes.FieldDataType) error {
-	if len(keys) == 0 {
-		return nil
-	}
+// 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
+//
+// 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",
+	})
 
-	var filteredKeys []*telemetrytypes.TelemetryFieldKey
-	for _, key := range keys {
-		if key.FieldContext == telemetrytypes.FieldContextBody {
-			filteredKeys = append(filteredKeys, key)
+	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")
 		}
+
+		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")
 	}
 
-	// collect paths for batch queries
-	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
 		}
 	}
@@ -94,6 +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_json_path_types 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_json_path_types 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_json_path_types 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_json_path_types 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_json_path_types 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

@@ -368,19 +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
-	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{
@@ -410,149 +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
-			}
-			if !strings.Contains(sel.Name, telemetrytypes.ArraySep) {
-				continue
-			}
-			key := &telemetrytypes.TelemetryFieldKey{
-				Name:         sel.Name,
-				Signal:       telemetrytypes.SignalLogs,
-				FieldContext: telemetrytypes.FieldContextBody,
-			}
-			for _, ps := range key.ArrayParentSelectors() {
-				parentPaths[ps.Name] = 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,
-			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,
-			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),
-			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("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(
 			"name 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)
+			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("name", 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("name", "%"+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("name", arm.dataTypeColumn)
-
 		query, args := sb.BuildWithFlavor(sqlbuilder.ClickHouse)
 		queries = append(queries, query)
 		allArgs = append(allArgs, args...)
@@ -589,7 +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{}
 
@@ -613,17 +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
@@ -677,11 +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.enrichBodyKeys(ctx, keys, parentTypeCache); err != nil {
-			t.logger.ErrorContext(ctx, "failed to enrich body JSON keys", errors.Attr(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

@@ -1,11 +1,13 @@
 package telemetrymetadata
 
+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                   = "distributed_field_keys"
+	PathTypesTableName               = otelcollectorconst.DistributedPathTypesTable
 	// 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/telemetrytraces/const.go

@@ -1,6 +1,50 @@
 package telemetrytraces
 
-import "github.com/SigNoz/signoz/pkg/types/telemetrytypes"
+import (
+	"github.com/SigNoz/signoz/pkg/types/telemetrytypes"
+)
+
+const (
+
+	// Internal Columns.
+	SpanTimestampBucketStartColumn = "ts_bucket_start"
+	SpanResourceFingerPrintColumn  = "resource_fingerprint"
+
+	// Intrinsic Columns.
+	SpanTimestampColumn        = "timestamp"
+	SpanTraceIDColumn          = "trace_id"
+	SpanSpanIDColumn           = "span_id"
+	SpanTraceStateColumn       = "trace_state"
+	SpanParentSpanIDColumn     = "parent_span_id"
+	SpanFlagsColumn            = "flags"
+	SpanNameColumn             = "name"
+	SpanKindColumn             = "kind"
+	SpanKindStringColumn       = "kind_string"
+	SpanDurationNanoColumn     = "duration_nano"
+	SpanStatusCodeColumn       = "status_code"
+	SpanStatusMessageColumn    = "status_message"
+	SpanStatusCodeStringColumn = "status_code_string"
+	SpanEventsColumn           = "events"
+	SpanLinksColumn            = "links"
+
+	// Calculated Columns.
+	SpanResponseStatusCodeColumn = "response_status_code"
+	SpanExternalHTTPURLColumn    = "external_http_url"
+	SpanHTTPURLColumn            = "http_url"
+	SpanExternalHTTPMethodColumn = "external_http_method"
+	SpanHTTPMethodColumn         = "http_method"
+	SpanHTTPHostColumn           = "http_host"
+	SpanDBNameColumn             = "db_name"
+	SpanDBOperationColumn        = "db_operation"
+	SpanHasErrorColumn           = "has_error"
+	SpanIsRemoteColumn           = "is_remote"
+
+	// Contextual Columns.
+	SpanAttributesStringColumn = "attributes_string"
+	SpanAttributesNumberColumn = "attributes_number"
+	SpanAttributesBoolColumn   = "attributes_bool"
+	SpanResourcesStringColumn  = "resources_string"
+)
 
 var (
 	IntrinsicFields = map[string]telemetrytypes.TelemetryFieldKey{

pkg/telemetrytraces/field_mapper_test.go

@@ -78,6 +78,16 @@ func TestGetFieldKeyName(t *testing.T) {
 			expectedResult: "multiIf(resource.`deployment.environment` IS NOT NULL, resource.`deployment.environment`::String, `resource_string_deployment$$environment_exists`==true, `resource_string_deployment$$environment`, NULL)",
 			expectedError:  nil,
 		},
+		{
+			name: "Contextual map column - attributes_string without span context does not short-circuit",
+			key: telemetrytypes.TelemetryFieldKey{
+				Name:          SpanAttributesStringColumn,
+				FieldContext:  telemetrytypes.FieldContextAttribute,
+				FieldDataType: telemetrytypes.FieldDataTypeString,
+			},
+			expectedResult: "attributes_string['attributes_string']",
+			expectedError:  nil,
+		},
 		{
 			name: "Non-existent column",
 			key: telemetrytypes.TelemetryFieldKey{

pkg/telemetrytraces/statement_builder.go

@@ -4,7 +4,6 @@ import (
 	"context"
 	"fmt"
 	"log/slog"
-	"slices"
 	"strings"
 
 	"github.com/SigNoz/signoz/pkg/errors"
@@ -15,7 +14,6 @@ import (
 	qbtypes "github.com/SigNoz/signoz/pkg/types/querybuildertypes/querybuildertypesv5"
 	"github.com/SigNoz/signoz/pkg/types/telemetrytypes"
 	"github.com/huandu/go-sqlbuilder"
-	"golang.org/x/exp/maps"
 )
 
 var (
@@ -86,40 +84,12 @@ func (b *traceQueryStatementBuilder) Build(
 		return nil, err
 	}
 
-	/*
-		Adding a tech debt note here:
-		This piece of code is a hot fix and should be removed once we close issue: engineering-pod/issues/3622
-	*/
-	/*
-		-------------------------------- Start of tech debt ----------------------------
-	*/
+	isSelectFieldsEmpty := false
 	if requestType == qbtypes.RequestTypeRaw {
-
-		selectedFields := query.SelectFields
-
-		if len(selectedFields) == 0 {
-			sortedKeys := maps.Keys(DefaultFields)
-			slices.Sort(sortedKeys)
-			for _, key := range sortedKeys {
-				selectedFields = append(selectedFields, DefaultFields[key])
-			}
-			query.SelectFields = selectedFields
-		}
-
-		selectFieldKeys := []string{}
-		for _, field := range selectedFields {
-			selectFieldKeys = append(selectFieldKeys, field.Name)
-		}
-
-		for _, x := range []string{"timestamp", "span_id", "trace_id"} {
-			if !slices.Contains(selectFieldKeys, x) {
-				query.SelectFields = append(query.SelectFields, DefaultFields[x])
-			}
-		}
+		// we are expanding here to ensure that all the conflicts are taken care in adjustKeys
+		// i.e if there is a conflict we strip away context of the key in adjustKeys
+		query, isSelectFieldsEmpty = b.expandRawSelectFields(query)
 	}
-	/*
-		-------------------------------- End of tech debt ----------------------------
-	*/
 
 	query = b.adjustKeys(ctx, keys, query, requestType)
 
@@ -128,7 +98,7 @@ func (b *traceQueryStatementBuilder) Build(
 
 	switch requestType {
 	case qbtypes.RequestTypeRaw:
-		return b.buildListQuery(ctx, q, query, start, end, keys, variables)
+		return b.buildListQuery(ctx, q, query, start, end, keys, variables, isSelectFieldsEmpty)
 	case qbtypes.RequestTypeTimeSeries:
 		return b.buildTimeSeriesQuery(ctx, q, query, start, end, keys, variables)
 	case qbtypes.RequestTypeScalar:
@@ -292,6 +262,7 @@ func (b *traceQueryStatementBuilder) buildListQuery(
 	start, end uint64,
 	keys map[string][]*telemetrytypes.TelemetryFieldKey,
 	variables map[string]qbtypes.VariableItem,
+	isSelectFieldsEmpty bool,
 ) (*qbtypes.Statement, error) {
 
 	var (
@@ -306,7 +277,6 @@ func (b *traceQueryStatementBuilder) buildListQuery(
 		cteArgs = append(cteArgs, args)
 	}
 
-	// TODO: should we deprecate `SelectFields` and return everything from a span like we do for logs?
 	for _, field := range query.SelectFields {
 		colExpr, err := b.fm.ColumnExpressionFor(ctx, start, end, &field, keys)
 		if err != nil {
@@ -315,6 +285,13 @@ func (b *traceQueryStatementBuilder) buildListQuery(
 		sb.SelectMore(colExpr)
 	}
 
+	if isSelectFieldsEmpty {
+		sb.SelectMore(SpanAttributesStringColumn)
+		sb.SelectMore(SpanAttributesNumberColumn)
+		sb.SelectMore(SpanAttributesBoolColumn)
+		sb.SelectMore(SpanResourcesStringColumn)
+	}
+
 	// From table
 	sb.From(fmt.Sprintf("%s.%s", DBName, SpanIndexV3TableName))
 
@@ -838,3 +815,52 @@ func (b *traceQueryStatementBuilder) buildResourceFilterCTE(
 		variables,
 	)
 }
+
+// expandRawSelectFields populates SelectFields for raw (list view) queries.
+// It must be called before adjustKeys so that normalization runs over the full set.
+// Returns the updated query and whether the original SelectFields was empty (i.e. full expansion was performed).
+func (b *traceQueryStatementBuilder) expandRawSelectFields(query qbtypes.QueryBuilderQuery[qbtypes.TraceAggregation]) (qbtypes.QueryBuilderQuery[qbtypes.TraceAggregation], bool) {
+	wasEmpty := len(query.SelectFields) == 0
+	selectFields := []telemetrytypes.TelemetryFieldKey{
+		{Name: SpanTimestampColumn, FieldContext: telemetrytypes.FieldContextSpan},
+		{Name: SpanTraceIDColumn, FieldContext: telemetrytypes.FieldContextSpan},
+		{Name: SpanSpanIDColumn, FieldContext: telemetrytypes.FieldContextSpan},
+	}
+	if wasEmpty {
+		// Select all intrinsic columns
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanTraceStateColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanParentSpanIDColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanFlagsColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanNameColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanKindColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanKindStringColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanDurationNanoColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanStatusCodeColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanStatusMessageColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanStatusCodeStringColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanEventsColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanLinksColumn, FieldContext: telemetrytypes.FieldContextSpan})
+
+		// select all calculated columns
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanResponseStatusCodeColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanExternalHTTPURLColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanHTTPURLColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanExternalHTTPMethodColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanHTTPMethodColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanHTTPHostColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanDBNameColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanDBOperationColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanHasErrorColumn, FieldContext: telemetrytypes.FieldContextSpan})
+		selectFields = append(selectFields, telemetrytypes.TelemetryFieldKey{Name: SpanIsRemoteColumn, FieldContext: telemetrytypes.FieldContextSpan})
+	} else {
+		for _, field := range query.SelectFields {
+			// TODO(tvats): If a user specifies attribute.timestamp in the select fields, this loop will basically ignore it, as we already added a field by default. This can be fixed once we close https://github.com/SigNoz/engineering-pod/issues/3693
+			if field.Name == SpanTimestampColumn || field.Name == SpanTraceIDColumn || field.Name == SpanSpanIDColumn {
+				continue
+			}
+			selectFields = append(selectFields, field)
+		}
+	}
+	query.SelectFields = selectFields
+	return query, wasEmpty
+}

pkg/telemetrytraces/stmt_builder_test.go

@@ -436,7 +436,7 @@ func TestStatementBuilderListQuery(t *testing.T) {
 				},
 			},
 			expected: qbtypes.Statement{
-				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE (simpleJSONExtractString(labels, 'service.name') = ? AND labels LIKE ? AND labels LIKE ?) AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT name AS `name`, multiIf(resource.`service.name` IS NOT NULL, resource.`service.name`::String, mapContains(resources_string, 'service.name'), resources_string['service.name'], NULL) AS `service.name`, duration_nano AS `duration_nano`, `attribute_number_cart$$items_count` AS `cart.items_count`, timestamp AS `timestamp`, span_id AS `span_id`, trace_id AS `trace_id` FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND true AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? LIMIT ?",
+				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE (simpleJSONExtractString(labels, 'service.name') = ? AND labels LIKE ? AND labels LIKE ?) AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp AS `timestamp`, trace_id AS `trace_id`, span_id AS `span_id`, name AS `name`, multiIf(resource.`service.name` IS NOT NULL, resource.`service.name`::String, mapContains(resources_string, 'service.name'), resources_string['service.name'], NULL) AS `service.name`, duration_nano AS `duration_nano`, `attribute_number_cart$$items_count` AS `cart.items_count` FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND true AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? LIMIT ?",
 				Args:  []any{"redis-manual", "%service.name%", "%service.name\":\"redis-manual%", uint64(1747945619), uint64(1747983448), "1747947419000000000", "1747983448000000000", uint64(1747945619), uint64(1747983448), 10},
 			},
 			expectedErr: nil,
@@ -465,7 +465,7 @@ func TestStatementBuilderListQuery(t *testing.T) {
 				Limit: 10,
 			},
 			expected: qbtypes.Statement{
-				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE (simpleJSONExtractString(labels, 'service.name') = ? AND labels LIKE ? AND labels LIKE ?) AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT duration_nano AS `duration_nano`, name AS `name`, response_status_code AS `response_status_code`, multiIf(resource.`service.name` IS NOT NULL, resource.`service.name`::String, mapContains(resources_string, 'service.name'), resources_string['service.name'], NULL) AS `service.name`, span_id AS `span_id`, timestamp AS `timestamp`, trace_id AS `trace_id` FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND true AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? ORDER BY attributes_string['user.id'] AS `user.id` desc LIMIT ?",
+				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE (simpleJSONExtractString(labels, 'service.name') = ? AND labels LIKE ? AND labels LIKE ?) AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp AS `timestamp`, trace_id AS `trace_id`, span_id AS `span_id`, trace_state AS `trace_state`, parent_span_id AS `parent_span_id`, flags AS `flags`, name AS `name`, kind AS `kind`, kind_string AS `kind_string`, duration_nano AS `duration_nano`, status_code AS `status_code`, status_message AS `status_message`, status_code_string AS `status_code_string`, events AS `events`, links AS `links`, response_status_code AS `response_status_code`, external_http_url AS `external_http_url`, http_url AS `http_url`, external_http_method AS `external_http_method`, http_method AS `http_method`, http_host AS `http_host`, db_name AS `db_name`, db_operation AS `db_operation`, has_error AS `has_error`, is_remote AS `is_remote`, attributes_string, attributes_number, attributes_bool, resources_string FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND true AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? ORDER BY attributes_string['user.id'] AS `user.id` desc LIMIT ?",
 				Args:  []any{"redis-manual", "%service.name%", "%service.name\":\"redis-manual%", uint64(1747945619), uint64(1747983448), "1747947419000000000", "1747983448000000000", uint64(1747945619), uint64(1747983448), 10},
 			},
 			expectedErr: nil,
@@ -509,7 +509,7 @@ func TestStatementBuilderListQuery(t *testing.T) {
 				Limit: 10,
 			},
 			expected: qbtypes.Statement{
-				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE (simpleJSONExtractString(labels, 'service.name') = ? AND labels LIKE ? AND labels LIKE ?) AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT name AS `name`, resource_string_service$$name AS `serviceName`, duration_nano AS `durationNano`, http_method AS `httpMethod`, response_status_code AS `responseStatusCode`, timestamp AS `timestamp`, span_id AS `span_id`, trace_id AS `trace_id` FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND true AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? LIMIT ?",
+				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE (simpleJSONExtractString(labels, 'service.name') = ? AND labels LIKE ? AND labels LIKE ?) AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp AS `timestamp`, trace_id AS `trace_id`, span_id AS `span_id`, name AS `name`, resource_string_service$$name AS `serviceName`, duration_nano AS `durationNano`, http_method AS `httpMethod`, response_status_code AS `responseStatusCode` FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND true AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? LIMIT ?",
 				Args:  []any{"redis-manual", "%service.name%", "%service.name\":\"redis-manual%", uint64(1747945619), uint64(1747983448), "1747947419000000000", "1747983448000000000", uint64(1747945619), uint64(1747983448), 10},
 			},
 			expectedErr: nil,
@@ -553,7 +553,7 @@ func TestStatementBuilderListQuery(t *testing.T) {
 				Limit: 10,
 			},
 			expected: qbtypes.Statement{
-				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE (simpleJSONExtractString(labels, 'service.name') = ? AND labels LIKE ? AND labels LIKE ?) AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT name AS `name`, resource_string_service$$name AS `serviceName`, duration_nano AS `durationNano`, http_method AS `httpMethod`, multiIf(toString(`attribute_string_mixed$$materialization$$key`) != '', toString(`attribute_string_mixed$$materialization$$key`), toString(multiIf(resource.`mixed.materialization.key` IS NOT NULL, resource.`mixed.materialization.key`::String, mapContains(resources_string, 'mixed.materialization.key'), resources_string['mixed.materialization.key'], NULL)) != '', toString(multiIf(resource.`mixed.materialization.key` IS NOT NULL, resource.`mixed.materialization.key`::String, mapContains(resources_string, 'mixed.materialization.key'), resources_string['mixed.materialization.key'], NULL)), NULL) AS `mixed.materialization.key`, timestamp AS `timestamp`, span_id AS `span_id`, trace_id AS `trace_id` FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND true AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? LIMIT ?",
+				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE (simpleJSONExtractString(labels, 'service.name') = ? AND labels LIKE ? AND labels LIKE ?) AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp AS `timestamp`, trace_id AS `trace_id`, span_id AS `span_id`, name AS `name`, resource_string_service$$name AS `serviceName`, duration_nano AS `durationNano`, http_method AS `httpMethod`, multiIf(toString(`attribute_string_mixed$$materialization$$key`) != '', toString(`attribute_string_mixed$$materialization$$key`), toString(multiIf(resource.`mixed.materialization.key` IS NOT NULL, resource.`mixed.materialization.key`::String, mapContains(resources_string, 'mixed.materialization.key'), resources_string['mixed.materialization.key'], NULL)) != '', toString(multiIf(resource.`mixed.materialization.key` IS NOT NULL, resource.`mixed.materialization.key`::String, mapContains(resources_string, 'mixed.materialization.key'), resources_string['mixed.materialization.key'], NULL)), NULL) AS `mixed.materialization.key` FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND true AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? LIMIT ?",
 				Args:  []any{"redis-manual", "%service.name%", "%service.name\":\"redis-manual%", uint64(1747945619), uint64(1747983448), "1747947419000000000", "1747983448000000000", uint64(1747945619), uint64(1747983448), 10},
 			},
 			expectedErr: nil,
@@ -598,7 +598,7 @@ func TestStatementBuilderListQuery(t *testing.T) {
 				Limit: 10,
 			},
 			expected: qbtypes.Statement{
-				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE (simpleJSONExtractString(labels, 'service.name') = ? AND labels LIKE ? AND labels LIKE ?) AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT name AS `name`, resource_string_service$$name AS `serviceName`, duration_nano AS `durationNano`, http_method AS `httpMethod`, `attribute_string_mixed$$materialization$$key` AS `mixed.materialization.key`, timestamp AS `timestamp`, span_id AS `span_id`, trace_id AS `trace_id` FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND true AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? LIMIT ?",
+				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE (simpleJSONExtractString(labels, 'service.name') = ? AND labels LIKE ? AND labels LIKE ?) AND seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp AS `timestamp`, trace_id AS `trace_id`, span_id AS `span_id`, name AS `name`, resource_string_service$$name AS `serviceName`, duration_nano AS `durationNano`, http_method AS `httpMethod`, `attribute_string_mixed$$materialization$$key` AS `mixed.materialization.key` FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND true AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? LIMIT ?",
 				Args:  []any{"redis-manual", "%service.name%", "%service.name\":\"redis-manual%", uint64(1747945619), uint64(1747983448), "1747947419000000000", "1747983448000000000", uint64(1747945619), uint64(1747983448), 10},
 			},
 			expectedErr: nil,
@@ -706,7 +706,7 @@ func TestStatementBuilderListQueryWithCorruptData(t *testing.T) {
 				Limit:        10,
 			},
 			expected: qbtypes.Statement{
-				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT duration_nano AS `duration_nano`, name AS `name`, response_status_code AS `response_status_code`, multiIf(resource.`service.name` IS NOT NULL, resource.`service.name`::String, mapContains(resources_string, 'service.name'), resources_string['service.name'], NULL) AS `service.name`, span_id AS `span_id`, timestamp AS `timestamp`, trace_id AS `trace_id` FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? LIMIT ?",
+				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp AS `timestamp`, trace_id AS `trace_id`, span_id AS `span_id`, trace_state AS `trace_state`, parent_span_id AS `parent_span_id`, flags AS `flags`, name AS `name`, kind AS `kind`, kind_string AS `kind_string`, duration_nano AS `duration_nano`, status_code AS `status_code`, status_message AS `status_message`, status_code_string AS `status_code_string`, events AS `events`, links AS `links`, response_status_code AS `response_status_code`, external_http_url AS `external_http_url`, http_url AS `http_url`, external_http_method AS `external_http_method`, http_method AS `http_method`, http_host AS `http_host`, db_name AS `db_name`, db_operation AS `db_operation`, has_error AS `has_error`, is_remote AS `is_remote`, attributes_string, attributes_number, attributes_bool, resources_string FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? LIMIT ?",
 				Args:  []any{uint64(1747945619), uint64(1747983448), "1747947419000000000", "1747983448000000000", uint64(1747945619), uint64(1747983448), 10},
 			},
 			expectedErr: nil,
@@ -739,7 +739,7 @@ func TestStatementBuilderListQueryWithCorruptData(t *testing.T) {
 				}},
 			},
 			expected: qbtypes.Statement{
-				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT duration_nano AS `duration_nano`, name AS `name`, response_status_code AS `response_status_code`, multiIf(resource.`service.name` IS NOT NULL, resource.`service.name`::String, mapContains(resources_string, 'service.name'), resources_string['service.name'], NULL) AS `service.name`, span_id AS `span_id`, timestamp AS `timestamp`, trace_id AS `trace_id` FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? ORDER BY timestamp AS `timestamp` asc LIMIT ?",
+				Query: "WITH __resource_filter AS (SELECT fingerprint FROM signoz_traces.distributed_traces_v3_resource WHERE seen_at_ts_bucket_start >= ? AND seen_at_ts_bucket_start <= ?) SELECT timestamp AS `timestamp`, trace_id AS `trace_id`, span_id AS `span_id`, trace_state AS `trace_state`, parent_span_id AS `parent_span_id`, flags AS `flags`, name AS `name`, kind AS `kind`, kind_string AS `kind_string`, duration_nano AS `duration_nano`, status_code AS `status_code`, status_message AS `status_message`, status_code_string AS `status_code_string`, events AS `events`, links AS `links`, response_status_code AS `response_status_code`, external_http_url AS `external_http_url`, http_url AS `http_url`, external_http_method AS `external_http_method`, http_method AS `http_method`, http_host AS `http_host`, db_name AS `db_name`, db_operation AS `db_operation`, has_error AS `has_error`, is_remote AS `is_remote`, attributes_string, attributes_number, attributes_bool, resources_string FROM signoz_traces.distributed_signoz_index_v3 WHERE resource_fingerprint GLOBAL IN (SELECT fingerprint FROM __resource_filter) AND timestamp >= ? AND timestamp < ? AND ts_bucket_start >= ? AND ts_bucket_start <= ? ORDER BY timestamp AS `timestamp` asc LIMIT ?",
 				Args:  []any{uint64(1747945619), uint64(1747983448), "1747947419000000000", "1747983448000000000", uint64(1747945619), uint64(1747983448), 10},
 			},
 			expectedErr: nil,

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
@@ -79,12 +80,6 @@ func (f *TelemetryFieldKey) ArrayParentSelectors() []*FieldKeySelector {
 	return selectors
 }
 
-// GetJSONDataType derives the JSONDataType from FieldDataType.
-// Callers should check FieldDataType != FieldDataTypeUnspecified before calling.
-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)
@@ -97,6 +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 {
@@ -119,6 +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("[]object")}
+	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

@@ -43,7 +43,7 @@ type JSONAccessNode struct {
 	isRoot     bool // marked true for only body_v2 and body_promoted
 
 	// Precomputed type information (single source of truth)
-	AvailableTypes []FieldDataType
+	AvailableTypes []JSONDataType
 
 	// Array type branches (Array(JSON) vs Array(Dynamic))
 	Branches map[JSONAccessBranchType]*JSONAccessNode
@@ -106,7 +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.
@@ -155,14 +155,14 @@ func (pb *planBuilder) buildPlan(index int, parent *JSONAccessNode, isDynArrChil
 		MaxDynamicPaths: maxPaths,
 	}
 
-	hasJSON := slices.Contains(node.AvailableTypes, FieldDataTypeArrayJSON)
-	hasDynamic := slices.Contains(node.AvailableTypes, FieldDataTypeArrayDynamic)
+	hasJSON := slices.Contains(node.AvailableTypes, ArrayJSON)
+	hasDynamic := slices.Contains(node.AvailableTypes, ArrayDynamic)
 
 	// Configure terminal if this is the last part
 	if isTerminal {
 		node.TerminalConfig = &TerminalConfig{
 			Key:      pb.key,
-			ElemType: pb.key.GetJSONDataType(),
+			ElemType: *pb.key.JSONDataType,
 		}
 	} else {
 		var err error
@@ -185,7 +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,
 	}
 }
 
@@ -242,7 +242,7 @@ 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
@@ -255,7 +255,7 @@ 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
@@ -276,7 +276,7 @@ func TestPlanJSON_BasicStructure(t *testing.T) {
 		},
 		{
 			name:         "Empty path returns error",
-			key:          makeKey("", FieldDataTypeString, false),
+			key:          makeKey("", String, false),
 			expectErr:    true,
 			expectedYAML: "",
 		},
@@ -431,7 +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,
@@ -450,7 +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,
@@ -493,7 +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,
@@ -666,12 +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{
@@ -692,7 +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,

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

tests/integration/src/querier/04_traces.py

@@ -490,25 +490,24 @@ def test_traces_list(
                     "name": "A",
                     "signal": "traces",
                     "disabled": False,
+                    "selectFields": [
+                        {"name": "span_id"},
+                        {"name": "span.timestamp"},
+                        {"name": "trace_id"},
+                    ],
                     "order": [{"key": {"name": "timestamp"}, "direction": "desc"}],
                     "limit": 1,
                 },
             },
             HTTPStatus.OK,
             lambda x: [
-                x[3].duration_nano,
-                x[3].name,
-                x[3].response_status_code,
-                x[3].service_name,
                 x[3].span_id,
                 format_timestamp(x[3].timestamp),
                 x[3].trace_id,
             ],  # type: Callable[[List[Traces]], List[Any]]
         ),
         # Case 2: order by attribute timestamp field which is there in attributes as well
-        # This should break but it doesn't because attribute.timestamp gets adjusted to timestamp
-        # because of default trace.timestamp gets added by default and bug in field mapper picks
-        # instrinsic field
+        # attribute.timestamp gets adjusted to span.timestamp
         pytest.param(
             {
                 "type": "builder_query",
@@ -516,6 +515,11 @@ def test_traces_list(
                     "name": "A",
                     "signal": "traces",
                     "disabled": False,
+                    "selectFields": [
+                        {"name": "span_id"},
+                        {"name": "span.timestamp"},
+                        {"name": "trace_id"},
+                    ],
                     "order": [
                         {"key": {"name": "attribute.timestamp"}, "direction": "desc"}
                     ],
@@ -524,10 +528,6 @@ def test_traces_list(
             },
             HTTPStatus.OK,
             lambda x: [
-                x[3].duration_nano,
-                x[3].name,
-                x[3].response_status_code,
-                x[3].service_name,
                 x[3].span_id,
                 format_timestamp(x[3].timestamp),
                 x[3].trace_id,
@@ -553,7 +553,7 @@ def test_traces_list(
             ],  # type: Callable[[List[Traces]], List[Any]]
         ),
         # Case 4: select attribute.timestamp with empty order by
-        # This doesn't return any data because of where_clause using aliased timestamp
+        # This returns the one span which has attribute.timestamp
         pytest.param(
             {
                 "type": "builder_query",
@@ -567,7 +567,11 @@ def test_traces_list(
                 },
             },
             HTTPStatus.OK,
-            lambda x: [],  # type: Callable[[List[Traces]], List[Any]]
+            lambda x: [
+                x[0].span_id,
+                format_timestamp(x[0].timestamp),
+                x[0].trace_id,
+            ],  # type: Callable[[List[Traces]], List[Any]]
         ),
         # Case 5: select timestamp with timestamp order by
         pytest.param(
@@ -706,6 +710,112 @@ def test_traces_list_with_corrupt_data(
                 assert data[key] == value
 
 
+@pytest.mark.parametrize(
+    "select_fields,status_code,expected_keys",
+    [
+        pytest.param(
+            [],
+            HTTPStatus.OK,
+            [
+                # all intrinsic column
+                "timestamp",
+                "trace_id",
+                "span_id",
+                "trace_state",
+                "parent_span_id",
+                "flags",
+                "name",
+                "kind",
+                "kind_string",
+                "duration_nano",
+                "status_code",
+                "status_message",
+                "status_code_string",
+                "events",
+                "links",
+                # all calculated columns
+                "response_status_code",
+                "external_http_url",
+                "http_url",
+                "external_http_method",
+                "http_method",
+                "http_host",
+                "db_name",
+                "db_operation",
+                "has_error",
+                "is_remote",
+                # all contextual columns (merged in response layer)
+                "attributes",
+                "resource",
+            ],
+        ),
+        pytest.param(
+            [
+                {"name": "service.name"},
+            ],
+            HTTPStatus.OK,
+            ["timestamp", "trace_id", "span_id", "service.name"],
+        ),
+    ],
+)
+def test_traces_list_with_select_fields(
+    signoz: types.SigNoz,
+    create_user_admin: None,  # pylint: disable=unused-argument
+    get_token: Callable[[str, str], str],
+    insert_traces: Callable[[List[Traces]], None],
+    select_fields: List[dict],
+    status_code: HTTPStatus,
+    expected_keys: List[str],
+) -> None:
+    """
+    Setup:
+    Insert 4 traces with different attributes.
+
+    Tests:
+    1. Empty select fields should return all the fields.
+    2. Non empty select field should return the select field along with timestamp, trace_id and span_id.
+    """
+    traces = (
+        generate_traces_with_corrupt_metadata()
+    )  # using this as the data doesn't matter
+
+    insert_traces(traces)
+
+    token = get_token(USER_ADMIN_EMAIL, USER_ADMIN_PASSWORD)
+
+    payload = {
+        "type": "builder_query",
+        "spec": {
+            "name": "A",
+            "signal": "traces",
+            "selectFields": select_fields,
+            "order": [{"key": {"name": "timestamp"}, "direction": "desc"}],
+            "limit": 1,
+        },
+    }
+
+    response = make_query_request(
+        signoz,
+        token,
+        start_ms=int(
+            (datetime.now(tz=timezone.utc) - timedelta(minutes=5)).timestamp() * 1000
+        ),
+        end_ms=int(datetime.now(tz=timezone.utc).timestamp() * 1000),
+        request_type="raw",
+        queries=[payload],
+    )
+    assert response.status_code == status_code
+
+    if response.status_code == HTTPStatus.OK:
+        data = response.json()
+        assert len(data["data"]["data"]["results"][0]["rows"][0]["data"].keys()) == len(
+            expected_keys
+        )
+        assert set(data["data"]["data"]["results"][0]["rows"][0]["data"].keys()) == set(
+            expected_keys
+        )
+
+
 @pytest.mark.parametrize(
     "order_by,aggregation_alias,expected_status",
     [
@@ -2131,9 +2241,9 @@ def test_traces_list_filter_by_trace_id(
     narrow_start_ms = int((now - timedelta(minutes=5)).timestamp() * 1000)
     narrow_rows = _query(narrow_start_ms, now_ms)
 
-    assert (
-        len(narrow_rows) == 1
-    ), f"Expected 1 span for trace_id filter (narrow window), got {len(narrow_rows)}"
+    assert len(narrow_rows) == 1, (
+        f"Expected 1 span for trace_id filter (narrow window), got {len(narrow_rows)}"
+    )
     assert narrow_rows[0]["data"]["span_id"] == span_id_root
     assert narrow_rows[0]["data"]["trace_id"] == target_trace_id