import dagster as dg from dagster_snowflake import SnowflakeResource from src.utils.snowflake.constants import Warehouse from src.utils.snowflake.query import load_query USER_DELETE_REQUEST_TABLE_NAME = "RDS_USER_DELETE_REQUEST" WAREHOUSE = Warehouse.SUNO_PROD_RDS_HOURLY_X_SMALL.value def parse_daily_partition_key(partition_key: str) -> str: """ Parse daily partition key to extract date. Args: partition_key: String in format "YYYY-MM-DD" Returns: String date in format "YYYY-MM-DD" """ return partition_key @dg.asset_check(asset="rds_user_delete_request", description="Verify ID uniqueness within partition", blocking=False) def check_user_delete_request_id_uniqueness(context: dg.AssetCheckExecutionContext, snowflake: SnowflakeResource) -> dg.AssetCheckResult: """ Checks for ID uniqueness within the current daily partition of the RDS_USER_DELETE_REQUEST table. """ partition_key = context.op_execution_context.partition_key p_date = parse_daily_partition_key(partition_key) with snowflake.get_connection() as conn: cursor = conn.cursor() cursor.execute(load_query("src/utils/snowflake/queries/use_warehouse.sql", params={"warehouse": WAREHOUSE})) query = f""" SELECT COUNT(*), COUNT(DISTINCT ID), COUNT(*) - COUNT(DISTINCT ID) FROM {USER_DELETE_REQUEST_TABLE_NAME} WHERE P_DATE = '{p_date}' """ cursor.execute(query) total, distinct, dupes = cursor.fetchone() or (0, 0, 0) return dg.AssetCheckResult( passed=dupes == 0, description=f"{'Unique' if dupes == 0 else f'{dupes} duplicates'}", metadata={ "total_rows": total, "distinct_ids": distinct, "duplicates": dupes, "partition_key": partition_key, } ) @dg.asset_check(asset="rds_user_delete_request", description="Verify row count > 0 for the partition", blocking=False) def check_user_delete_request_row_count(context: dg.AssetCheckExecutionContext, snowflake: SnowflakeResource) -> dg.AssetCheckResult: """ Checks that the row count for the current daily partition of the RDS_USER_DELETE_REQUEST table is greater than 0. """ partition_key = context.op_execution_context.partition_key p_date = parse_daily_partition_key(partition_key) with snowflake.get_connection() as conn: cursor = conn.cursor() cursor.execute(load_query("src/utils/snowflake/queries/use_warehouse.sql", params={"warehouse": WAREHOUSE})) query = f""" SELECT COUNT(*) FROM {USER_DELETE_REQUEST_TABLE_NAME} WHERE P_DATE = '{p_date}' """ cursor.execute(query) row_count = cursor.fetchone()[0] return dg.AssetCheckResult( passed=row_count > 0, description=f"Table has {row_count} rows for partition {partition_key}", metadata={ "row_count": row_count, "partition_key": partition_key, } ) asset_checks = [ check_user_delete_request_id_uniqueness, check_user_delete_request_row_count, ] __all__ = ["asset_checks"]