feat(ingest/presto-on-hive): enable partition key for presto-on-hive (#…

…8380)
datahub-project · Aug 2, 2023 · bf47d65 · bf47d65
1 parent 4f9fc67
commit bf47d65
Show file tree

Hide file tree

Showing 12 changed files with 51 additions and 0 deletions.
diff --git a/...e/src/main/java/com/linkedin/datahub/graphql/types/dataset/mappers/SchemaFieldMapper.java b/...e/src/main/java/com/linkedin/datahub/graphql/types/dataset/mappers/SchemaFieldMapper.java
@@ -34,6 +34,7 @@ public SchemaField apply(@Nonnull final com.linkedin.schema.SchemaField input, @
             result.setGlossaryTerms(GlossaryTermsMapper.map(input.getGlossaryTerms(), entityUrn));
         }
         result.setIsPartOfKey(input.isIsPartOfKey());
+        result.setIsPartitioningKey(input.isIsPartitioningKey());
         return result;
     }
 

diff --git a/datahub-graphql-core/src/main/resources/entity.graphql b/datahub-graphql-core/src/main/resources/entity.graphql
@@ -2872,6 +2872,11 @@ type SchemaField {
     Whether the field is part of a key schema
     """
     isPartOfKey: Boolean
+
+    """
+    Whether the field is part of a partitioning key schema
+    """
+    isPartitioningKey: Boolean
 }
 
 """

diff --git a/datahub-web-react/src/Mocks.tsx b/datahub-web-react/src/Mocks.tsx
@@ -549,6 +549,7 @@ export const dataset3WithSchema = {
                     type: SchemaFieldDataType.String,
                     nativeDataType: 'varchar(100)',
                     isPartOfKey: false,
+                    isPartitioningKey: false,
                     jsonPath: null,
                     globalTags: null,
                     glossaryTerms: null,
@@ -563,6 +564,7 @@ export const dataset3WithSchema = {
                     type: SchemaFieldDataType.String,
                     nativeDataType: 'boolean',
                     isPartOfKey: false,
+                    isPartitioningKey: false,
                     jsonPath: null,
                     globalTags: null,
                     glossaryTerms: null,

diff --git a/datahub-web-react/src/app/entity/dataset/profile/schema/utils/schemaTitleRenderer.tsx b/datahub-web-react/src/app/entity/dataset/profile/schema/utils/schemaTitleRenderer.tsx
@@ -7,6 +7,7 @@ import { ExtendedSchemaFields } from './types';
 import TypeLabel from '../../../../shared/tabs/Dataset/Schema/components/TypeLabel';
 import { ForeignKeyConstraint, SchemaMetadata } from '../../../../../../types.generated';
 import PrimaryKeyLabel from '../../../../shared/tabs/Dataset/Schema/components/PrimaryKeyLabel';
+import PartitioningKeyLabel from '../../../../shared/tabs/Dataset/Schema/components/PartitioningKeyLabel';
 import NullableLabel from '../../../../shared/tabs/Dataset/Schema/components/NullableLabel';
 import ForeignKeyLabel from '../../../../shared/tabs/Dataset/Schema/components/ForeignKeyLabel';
 
@@ -62,6 +63,7 @@ export default function useSchemaTitleRenderer(
                     </FieldPathText>
                     <TypeLabel type={record.type} nativeDataType={record.nativeDataType} />
                     {(schemaMetadata?.primaryKeys?.includes(fieldPath) || record.isPartOfKey) && <PrimaryKeyLabel />}
+                    {record.isPartitioningKey && <PartitioningKeyLabel />}
                     {record.nullable && <NullableLabel />}
                     {schemaMetadata?.foreignKeys
                         ?.filter(

diff --git a/...b-web-react/src/app/entity/shared/tabs/Dataset/Schema/components/PartitioningKeyLabel.tsx b/...b-web-react/src/app/entity/shared/tabs/Dataset/Schema/components/PartitioningKeyLabel.tsx
@@ -0,0 +1,21 @@
+import React from 'react';
+import { Badge } from 'antd';
+import styled from 'styled-components';
+import { blue } from '@ant-design/colors';
+import { ANTD_GRAY } from '../../../../constants';
+
+const PartitioningKeyBadge = styled(Badge)`
+    margin-left: 4px;
+    &&& .ant-badge-count {
+        background-color: ${ANTD_GRAY[1]};
+        color: ${blue[5]};
+        border: 1px solid ${blue[2]};
+        font-size: 12px;
+        font-weight: 400;
+        height: 22px;
+    }
+`;
+
+export default function PartitioningKeyLabel() {
+    return <PartitioningKeyBadge count="Partition Key" />;
+}
diff --git a/datahub-web-react/src/graphql/fragments.graphql b/datahub-web-react/src/graphql/fragments.graphql
@@ -678,6 +678,7 @@ fragment schemaFieldFields on SchemaField {
     nativeDataType
     recursive
     isPartOfKey
+    isPartitioningKey
     globalTags {
         ...globalTagsFields
     }

diff --git a/datahub-web-react/src/graphql/versionedDataset.graphql b/datahub-web-react/src/graphql/versionedDataset.graphql
@@ -10,6 +10,7 @@ query getVersionedDataset($urn: String!, $versionStamp: String) {
                 nativeDataType
                 recursive
                 isPartOfKey
+                isPartitioningKey
             }
             lastObserved
         }

diff --git a/metadata-ingestion/src/datahub/ingestion/source/sql/presto_on_hive.py b/metadata-ingestion/src/datahub/ingestion/source/sql/presto_on_hive.py
@@ -524,6 +524,8 @@ def loop_tables(
             # add table schema fields
             schema_fields = self.get_schema_fields(dataset_name, columns)
 
+            self._set_partition_key(columns, schema_fields)
+
             schema_metadata = get_schema_metadata(
                 self.report,
                 dataset_name,
@@ -888,6 +890,18 @@ def get_schema_fields_for_column(
             default_nullable=True,
         )
 
+    def _set_partition_key(self, columns, schema_fields):
+        if len(columns) > 0:
+            partition_key_names = set()
+            for column in columns:
+                if column["is_partition_col"]:
+                    partition_key_names.add(column["col_name"])
+
+            for schema_field in schema_fields:
+                name = schema_field.fieldPath.split(".")[-1]
+                if name in partition_key_names:
+                    schema_field.isPartitioningKey = True
+
 
 class SQLAlchemyClient:
     def __init__(self, config: SQLAlchemyConfig):

diff --git a/metadata-ingestion/tests/integration/presto-on-hive/presto_on_hive_mces_golden_1.json b/metadata-ingestion/tests/integration/presto-on-hive/presto_on_hive_mces_golden_1.json
@@ -1277,6 +1277,7 @@
                                 "nativeDataType": "string",
                                 "recursive": false,
                                 "isPartOfKey": false,
+                                "isPartitioningKey": true,
                                 "jsonProps": "{\"native_data_type\": \"string\", \"_nullable\": true}"
                             },
                             {

diff --git a/metadata-ingestion/tests/integration/presto-on-hive/presto_on_hive_mces_golden_2.json b/metadata-ingestion/tests/integration/presto-on-hive/presto_on_hive_mces_golden_2.json
@@ -1277,6 +1277,7 @@
                                 "nativeDataType": "string",
                                 "recursive": false,
                                 "isPartOfKey": false,
+                                "isPartitioningKey": true,
                                 "jsonProps": "{\"native_data_type\": \"string\", \"_nullable\": true}"
                             },
                             {

diff --git a/metadata-ingestion/tests/integration/presto-on-hive/presto_on_hive_mces_golden_3.json b/metadata-ingestion/tests/integration/presto-on-hive/presto_on_hive_mces_golden_3.json
@@ -1277,6 +1277,7 @@
                                 "nativeDataType": "string",
                                 "recursive": false,
                                 "isPartOfKey": false,
+                                "isPartitioningKey": true,
                                 "jsonProps": "{\"native_data_type\": \"string\", \"_nullable\": true}"
                             },
                             {

diff --git a/metadata-ingestion/tests/integration/presto-on-hive/presto_on_hive_mces_golden_4.json b/metadata-ingestion/tests/integration/presto-on-hive/presto_on_hive_mces_golden_4.json
@@ -1277,6 +1277,7 @@
                                 "nativeDataType": "string",
                                 "recursive": false,
                                 "isPartOfKey": false,
+                                "isPartitioningKey": true,
                                 "jsonProps": "{\"native_data_type\": \"string\", \"_nullable\": true}"
                             },
                             {