Add additional columns to housing_gx_data_quality_tests_complete and adjust configurations which include:

- generate table specific id columns using result.unexpected_index_list so that specific records can be selected, rather than indices
- a column that specifies the type of DQ dimension that a test is
- adjust results format so that partial counts are not returned
- upgrade GX to 1.2.0
- remove dtype mapping for partial counts
- change tenure_id to tenancy_id in housing_gx_dq_inputs.py

Author: annajgibson

scripts/helpers/housing_gx_dq_inputs.py

@@ -4,7 +4,7 @@
     'id_field': 'person_id'},
     'tenure_reshape': {
         'sql': """SELECT * FROM "housing-refined-zone"."tenure_reshape" where import_date>'20240412' and import_date=(select max(import_date) from "housing-refined-zone"."tenure_reshape" where import_date>'20240412') and isterminated=False and description in ('Secure', 'Introductory')""",
-        'id_field': 'tenure_id'},
+        'id_field': 'tenancy_id'},
     'contacts_reshape': {
         'sql': """SELECT id, targetid, createdat, contacttype, subtype, value, lastmodified, targettype, isactive, person_id, import_date  FROM "housing-refined-zone"."contacts_reshape"  where import_date=(select max(import_date) from "housing-refined-zone"."contacts_reshape") and isactive=True""",
         'id_field': 'id'},

scripts/jobs/housing/housing_apply_gx_dq_tests.py

@@ -10,7 +10,7 @@
 import great_expectations as gx
 import pandas as pd
 from pyathena import connect
-from scripts.helpers.housing_gx_dq_inputs import sql_config, table_list, partition_keys
+from scripts.helpers.housing_gx_dq_inputs import sql_config, table_list, partition_keys, dq_dimensions_map
 import scripts.jobs.housing.housing_person_reshape_gx_suite
 import scripts.jobs.housing.housing_tenure_reshape_gx_suite
 import scripts.jobs.housing.housing_contacts_reshape_gx_suite
@@ -98,7 +98,9 @@ def main():
                 name=f'{table}_checkpoint',
                 validation_definitions=[validation_definition],
                 actions=actions,
-                result_format={"result_format": "COMPLETE"}
+                result_format={"result_format": "COMPLETE",
+                               "return_unexpected_index_query": False,
+                               "partial_unexpected_count": 0}
             )
         )
 
@@ -111,8 +113,20 @@ def main():
         table_results_df = pd.json_normalize(results['validation_results'][0]['expectations'])
         table_results_df_list.append(table_results_df)
 
+        # generate id lists for each unexpected result set
+        query_df = table_results_df.loc[(~table_results_df['result.unexpected_index_list'].isna()) & (
+                table_results_df['result.unexpected_index_list'].values != '[]')]
+
+        table_results_df['unexpected_id_list'] = pd.Series(dtype='object')
+        for i, row in query_df.iterrows():
+            table_results_df.loc[i, 'unexpected_id_list'] = str(
+                list(df[sql_config.get(table).get('id_field')].iloc[row['result.unexpected_index_list']]))
+
     results_df = pd.concat(table_results_df_list)
 
+    # map DQ dimension type
+    results_df['dq_dimension_type'] = results_df['expectation_type'].map(dq_dimensions_map)
+
     results_df['import_year'] = datetime.today().year
     results_df['import_month'] = datetime.today().month
     results_df['import_day'] = datetime.today().day
@@ -128,8 +142,6 @@ def main():
                   'result.unexpected_count': 'bigint',
                   'result.missing_count': 'bigint',
                   'result.partial_unexpected_list': 'array<string>',
-                  'result.partial_unexpected_counts': 'array<struct<count:bigint,value:string>>',
-                  'result.partial_unexpected_index_list': 'array<bigint>',
                   'result.unexpected_list': 'array<string>',
                   'result.unexpected_index_list': 'array<bigint>',
                   'result.unexpected_index_query': 'string',

terraform/etl/54-aws-glue-housing-apply-gx-dq-tests.tf

@@ -20,7 +20,7 @@ module "housing_apply_gx_dq_tests" {
     "--job-bookmark-option"              = "job-bookmark-enable"
     "--enable-glue-datacatalog"          = "true"
     "--enable-continuous-cloudwatch-log" = "true"
-    "--additional-python-modules"        = "great_expectations==1.1.0,PyAthena,awswrangler"
+    "--additional-python-modules"        = "great_expectations==1.2.0,PyAthena,awswrangler"
     "--region_name"                      = data.aws_region.current.name
     "--s3_endpoint"                      = "https://s3.${data.aws_region.current.name}.amazonaws.com"
     "--s3_target_location"               = "s3://${module.raw_zone_data_source.bucket_id}/housing/data-quality-tests/"