fix(dq): seed NUMERIC on the numeric column test definitions

NUMERIC is a distinct member of the column dataType enum and is what BigQuery,
Postgres, Snowflake and DB2 numeric columns are ingested as, but the numeric
system test definitions only ever listed NUMBER and DECIMAL. The "Add test case"
dropdown filters on the column's exact dataType
(`TestCaseFormBody` -> `supportedDataType=<column.dataType>`), so on a NUMERIC
column the mean/min/max/median/stddev/sum/between tests were not offered at all —
only the three definitions that happened to list NUMERIC
(missingCount, notNull, unique) showed up.

- Seeds: add NUMERIC to the ten column definitions that list NUMBER but not
  NUMERIC. The three string-only definitions are deliberately untouched.
- Migration (2.1.0, MySQL + Postgres): backfill the same ten definitions.
  `initializeEntity` returns early when the entity exists, so seeding alone only
  fixes fresh installs. Guarded on NUMERIC being absent, so re-runs are a no-op,
  and on the array existing, so a definition with no supportedDataTypes — which
  already means "every data type" — is not narrowed to exactly one.
- IT: assert every seeded numeric aggregate definition is returned for NUMERIC,
  and that a string-only definition still is not.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

Author: TeddyCr

bootstrap/sql/migrations/native/2.1.0/mysql/postDataMigrationSQLScript.sql

@@ -236,6 +236,24 @@ WHERE name IN (
     '"dimensionFailurePolicy"'
   );
 
+-- NUMERIC is a distinct member of the column dataType enum and is what BigQuery, Postgres,
+-- Snowflake and DB2 numeric columns are ingested as, but the numeric system test definitions were
+-- only ever seeded with NUMBER/DECIMAL. The "Add test case" dropdown filters on the column's exact
+-- dataType, so mean/min/max/median/stddev/sum were unreachable on any NUMERIC column. Seeding only
+-- covers fresh installs (initializeEntity returns early when the entity exists), hence this
+-- backfill. The guard on NUMERIC being absent keeps re-runs a no-op, and it also skips a definition
+-- with no supportedDataTypes at all -- that already means "every data type" (issue #27718), so
+-- appending to it would narrow it to exactly one.
+UPDATE test_definition
+SET json = JSON_ARRAY_APPEND(json, '$.supportedDataTypes', 'NUMERIC')
+WHERE name IN (
+    'columnValueMaxToBeBetween', 'columnValueMeanToBeBetween', 'columnValueMedianToBeBetween',
+    'columnValueMinToBeBetween', 'columnValueStdDevToBeBetween',
+    'columnValuesToBeAtExpectedLocation', 'columnValuesSumToBeBetween', 'columnValuesToBeBetween',
+    'columnValuesToBeInSet', 'columnValuesToBeNotInSet'
+  )
+  AND NOT JSON_CONTAINS(json, JSON_QUOTE('NUMERIC'), '$.supportedDataTypes');
+
 -- Normalize user emails to lowercase: email is the primary identity lookup key and the
 -- application always compares lowercased values. The case-insensitive unique key on email
 -- guarantees no collisions can result from lowercasing.

bootstrap/sql/migrations/native/2.1.0/postgres/postDataMigrationSQLScript.sql

@@ -214,6 +214,29 @@ WHERE name IN (
   )
   AND NOT ((json->'parameterDefinition')::jsonb @> '[{"name": "dimensionFailurePolicy"}]'::jsonb);
 
+-- NUMERIC is a distinct member of the column dataType enum and is what BigQuery, Postgres,
+-- Snowflake and DB2 numeric columns are ingested as, but the numeric system test definitions were
+-- only ever seeded with NUMBER/DECIMAL. The "Add test case" dropdown filters on the column's exact
+-- dataType, so mean/min/max/median/stddev/sum were unreachable on any NUMERIC column. Seeding only
+-- covers fresh installs (initializeEntity returns early when the entity exists), hence this
+-- backfill. The guard on NUMERIC being absent keeps re-runs a no-op, and it also skips a definition
+-- with no supportedDataTypes at all -- that already means "every data type" (issue #27718), so
+-- appending to it would narrow it to exactly one.
+UPDATE test_definition
+SET json = jsonb_set(
+    json::jsonb,
+    '{supportedDataTypes}',
+    (json->'supportedDataTypes')::jsonb || '["NUMERIC"]'::jsonb
+)
+WHERE name IN (
+    'columnValueMaxToBeBetween', 'columnValueMeanToBeBetween', 'columnValueMedianToBeBetween',
+    'columnValueMinToBeBetween', 'columnValueStdDevToBeBetween',
+    'columnValuesToBeAtExpectedLocation', 'columnValuesSumToBeBetween', 'columnValuesToBeBetween',
+    'columnValuesToBeInSet', 'columnValuesToBeNotInSet'
+  )
+  AND json->'supportedDataTypes' IS NOT NULL
+  AND NOT ((json->'supportedDataTypes')::jsonb @> '["NUMERIC"]'::jsonb);
+
 -- Normalize user emails to lowercase: email is the primary identity lookup key and the
 -- application always compares lowercased values. No collision guard is needed -- the 1.5.0
 -- migration already deleted rows duplicated by LOWER(email) and lowercased the survivors, and

openmetadata-integration-tests/src/test/java/org/openmetadata/it/tests/TestDefinitionResourceIT.java

@@ -43,6 +43,15 @@ public class TestDefinitionResourceIT extends BaseEntityIT<TestDefinition, Creat
       List.of("COLUMN", "Column", "column", " Column ");
   private static final List<String> BLANK_ENTITY_TYPES = List.of("", " ");
   private static final int ENTITY_TYPE_FILTER_LIMIT = 1000000;
+  private static final List<String> NUMERIC_AGGREGATE_DEFINITIONS =
+      List.of(
+          "columnValueMaxToBeBetween",
+          "columnValueMeanToBeBetween",
+          "columnValueMedianToBeBetween",
+          "columnValueMinToBeBetween",
+          "columnValueStdDevToBeBetween",
+          "columnValuesSumToBeBetween");
+  private static final String REGEX_DEFINITION = "columnValuesToMatchRegex";
 
   // Disable tests that don't apply to TestDefinition
   {
@@ -363,6 +372,33 @@ void list_supportedDataTypeFilterKeepsDefinitionsWithoutDataTypes_200_OK(TestNam
             + " separate listCount query");
   }
 
+  /**
+   * NUMERIC is what BigQuery, Postgres, Snowflake and DB2 numeric columns are ingested as, but the
+   * seeded aggregate definitions only ever listed NUMBER and DECIMAL. Since the listing filters on
+   * the column's exact data type, none of mean/min/max/median/stddev/sum could be picked for a
+   * NUMERIC column.
+   */
+  @Test
+  void list_supportedDataTypeNumericReturnsSeededAggregateDefinitions_200_OK() {
+    Set<String> fullyQualifiedNames =
+        fullyQualifiedNamesOf(
+            listBySupportedDataType(SdkClients.adminClient(), ColumnDataType.NUMERIC.value())
+                .getData());
+
+    assertTrue(
+        fullyQualifiedNames.containsAll(NUMERIC_AGGREGATE_DEFINITIONS),
+        () ->
+            "Every seeded numeric aggregate definition must be offered for a NUMERIC column, missing: "
+                + NUMERIC_AGGREGATE_DEFINITIONS.stream()
+                    .filter(name -> !fullyQualifiedNames.contains(name))
+                    .collect(Collectors.joining(", ")));
+    assertFalse(
+        fullyQualifiedNames.contains(REGEX_DEFINITION),
+        REGEX_DEFINITION
+            + " only supports string types, so returning it for NUMERIC would mean the filter"
+            + " stopped discriminating rather than that the data types were corrected");
+  }
+
   private static ListResponse<TestDefinition> listBySupportedDataType(
       OpenMetadataClient client, String supportedDataType) {
     ListParams params =

openmetadata-service/src/main/resources/json/data/tests/columnValueMaxToBeBetween.json

@@ -5,7 +5,7 @@
     "description": "This schema defines the test ColumnValueMaxToBeBetween. Test the maximum value in a col is within a range.",
     "entityType": "COLUMN",
     "testPlatforms": ["OpenMetadata", "dbt"],
-    "supportedDataTypes": ["NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT"],
+    "supportedDataTypes": ["NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "NUMERIC", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT"],
     "parameterDefinition": [
     {
       "name": "minValueForMaxInCol",

openmetadata-service/src/main/resources/json/data/tests/columnValueMeanToBeBetween.json

@@ -5,7 +5,7 @@
     "description": "This schema defines the test ColumnValueMeanToBeBetween. Test the mean value in a col is within a range.",
     "entityType": "COLUMN",
     "testPlatforms": ["OpenMetadata"],
-    "supportedDataTypes": ["NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT", "ARRAY", "SET"],
+    "supportedDataTypes": ["NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "NUMERIC", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT", "ARRAY", "SET"],
     "parameterDefinition": [
     {
       "name": "minValueForMeanInCol",

openmetadata-service/src/main/resources/json/data/tests/columnValueMedianToBeBetween.json

@@ -5,7 +5,7 @@
     "description": "This schema defines the test ColumnValueMedianToBeBetween. Test the median value in a col is within a range.",
     "entityType": "COLUMN",
     "testPlatforms": ["OpenMetadata"],
-    "supportedDataTypes": ["NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT"],
+    "supportedDataTypes": ["NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "NUMERIC", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT"],
     "parameterDefinition": [
       {
         "name": "minValueForMedianInCol",

openmetadata-service/src/main/resources/json/data/tests/columnValueMinToBeBetween.json

@@ -5,7 +5,7 @@
     "description": "This schema defines the test ColumnValueMinToBeBetween. Test the minimum value in a col is within a range.",
     "entityType": "COLUMN",
     "testPlatforms": ["OpenMetadata"],
-    "supportedDataTypes": ["NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT"],
+    "supportedDataTypes": ["NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "NUMERIC", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT"],
     "parameterDefinition": [
       {
         "name": "minValueForMinInCol",

openmetadata-service/src/main/resources/json/data/tests/columnValueStdDevToBeBetween.json

@@ -5,7 +5,7 @@
     "description": "This schema defines the test ColumnValueStdDevToBeBetween. Test the std. dev. value in a col is within a range.",
     "entityType": "COLUMN",
     "testPlatforms": ["OpenMetadata"],
-    "supportedDataTypes": ["NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT"],
+    "supportedDataTypes": ["NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "NUMERIC", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT"],
     "parameterDefinition": [
       {
         "name": "minValueForStdDevInCol",

openmetadata-service/src/main/resources/json/data/tests/columnValueToBeAtExpectedLocation.json

@@ -5,7 +5,7 @@
     "description": "This schema defines the test ColumnValuesToBeAtExpectedLocation. Test the lat/long values in a column to be at the specified location in the reference column.",
     "entityType": "COLUMN",
     "testPlatforms": ["OpenMetadata"],
-    "supportedDataTypes": ["BYTES", "STRING", "MEDIUMTEXT", "TEXT", "CHAR", "VARCHAR","NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT"],
+    "supportedDataTypes": ["BYTES", "STRING", "MEDIUMTEXT", "TEXT", "CHAR", "VARCHAR","NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "NUMERIC", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT"],
     "parameterDefinition": [
       {
         "name": "locationReferenceType",

openmetadata-service/src/main/resources/json/data/tests/columnValuesSumToBeBetween.json

@@ -5,7 +5,7 @@
     "description": "This schema defines the test ColumnValuesSumToBeBetween. Test the sum of the values of a col is within a range.",
     "entityType": "COLUMN",
     "testPlatforms": ["OpenMetadata"],
-    "supportedDataTypes": ["NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT"],
+    "supportedDataTypes": ["NUMBER", "INT", "FLOAT", "DOUBLE", "DECIMAL", "NUMERIC", "TINYINT", "SMALLINT", "BIGINT", "BYTEINT"],
     "parameterDefinition": [
       {
         "name": "minValueForColSum",