From 427a06cfa82d91fde18a143a728fee5badfab697 Mon Sep 17 00:00:00 2001
From: Ayush Shah <ayush@getcollate.io>
Date: Sat, 18 Jan 2025 19:41:30 +0530
Subject: [PATCH] MINOR: BigQuery Improvement, Hive Partitioned Tables,
 Nonetype issue resolved (#19429)

---
 .../source/database/bigquery/metadata.py      | 25 +++++++++++++++++++
 .../database/bigquery/profiler/profiler.py    |  1 +
 ingestion/src/metadata/sampler/partition.py   |  3 ++-
 3 files changed, 28 insertions(+), 1 deletion(-)

diff --git a/ingestion/src/metadata/ingestion/source/database/bigquery/metadata.py b/ingestion/src/metadata/ingestion/source/database/bigquery/metadata.py
index 078853aca59..310cb203601 100644
--- a/ingestion/src/metadata/ingestion/source/database/bigquery/metadata.py
+++ b/ingestion/src/metadata/ingestion/source/database/bigquery/metadata.py
@@ -667,6 +667,31 @@ class BigquerySource(LifeCycleQueryMixin, CommonDbSourceService, MultiDBSource):
             database = self.context.get().database
             table = self.client.get_table(fqn._build(database, schema_name, table_name))
             columns = inspector.get_columns(table_name, schema_name, db_name=database)
+            if hasattr(table, "external_data_configuration") and hasattr(
+                table.external_data_configuration, "hive_partitioning"
+            ):
+                # Ingesting External Hive Partitioned Tables
+                from google.cloud.bigquery.external_config import (  # pylint: disable=import-outside-toplevel
+                    HivePartitioningOptions,
+                )
+
+                partition_details: HivePartitioningOptions = (
+                    table.external_data_configuration.hive_partitioning
+                )
+                return True, TablePartition(
+                    columns=[
+                        PartitionColumnDetails(
+                            columnName=self._get_partition_column_name(
+                                columns=columns,
+                                partition_field_name=field,
+                            ),
+                            interval=str(partition_details._properties.get("mode")),
+                            intervalType=PartitionIntervalTypes.OTHER,
+                        )
+                        for field in partition_details._properties.get("fields")
+                    ]
+                )
+
             if table.time_partitioning is not None:
                 if table.time_partitioning.field:
                     table_partition = TablePartition(
diff --git a/ingestion/src/metadata/ingestion/source/database/bigquery/profiler/profiler.py b/ingestion/src/metadata/ingestion/source/database/bigquery/profiler/profiler.py
index 2cd0f225b31..750ae61fb20 100644
--- a/ingestion/src/metadata/ingestion/source/database/bigquery/profiler/profiler.py
+++ b/ingestion/src/metadata/ingestion/source/database/bigquery/profiler/profiler.py
@@ -24,6 +24,7 @@ class BigQueryProfiler(BigQueryProfilerInterface):
         return self.system_metrics_computer.get_system_metrics(
             table=runner.dataset,
             usage_location=self.service_connection_config.usageLocation,
+            runner=runner,
         )
 
     def initialize_system_metrics_computer(self) -> BigQuerySystemMetricsComputer:
diff --git a/ingestion/src/metadata/sampler/partition.py b/ingestion/src/metadata/sampler/partition.py
index e62ad94a049..c2eb76f9833 100644
--- a/ingestion/src/metadata/sampler/partition.py
+++ b/ingestion/src/metadata/sampler/partition.py
@@ -49,7 +49,7 @@ def validate_athena_injected_partitioning(
 
     column_partitions: Optional[List[PartitionColumnDetails]] = table_partitions.columns
     if not column_partitions:
-        raise RuntimeError("Table parition is set but no columns are defined.")
+        raise RuntimeError("Table partition is set but no columns are defined.")
 
     for column_partition in column_partitions:
         if column_partition.intervalType == PartitionIntervalTypes.INJECTED:
@@ -163,6 +163,7 @@ def _handle_bigquery_partition(
                 partitionIntegerRangeStart=1,
                 partitionIntegerRangeEnd=10000,
             )
+        # TODO: Allow External Hive Partitioning for profiler
         raise TypeError(
             f"Unsupported partition type {partition.intervalType}. Skipping table"
         )