sws-spark-dissemination-helper 0.0.93__py3-none-any.whl → 0.0.95__py3-none-any.whl

sws_spark_dissemination_helper/SWSBronzeIcebergSparkHelper.py

@@ -447,6 +447,9 @@ class SWSBronzeIcebergSparkHelper:
     def write_bronze_disseminated_tag_data_to_iceberg_and_csv(
         self, dimensions: Dict[str, List[str]]
     ) -> DataFrame:
+        
+        refs = self.spark.sql(f"SELECT * FROM {self.iceberg_tables.BRONZE.iceberg_id}.refs").collect()
+        logging.info(f"bronze refs: {refs}")
 
         create_branch_query = f"ALTER TABLE {self.iceberg_tables.BRONZE.iceberg_id}.`tag_{self.tag_name}` CREATE OR REPLACE BRANCH `diss_tag_{self.tag_name}`"  # AS OF VERSION `{tag_name}`
         logging.info(f"create_branch_query: {create_branch_query}")
@@ -454,25 +457,36 @@ class SWSBronzeIcebergSparkHelper:
 
         logging.info(f"result of create_branch_query: {create_branch_query_result}")
 
+        self.disseminated_tag_df = self.spark.read.option(
+            "branch", f"`diss_tag_{self.tag_name}`"
+        ).table(self.iceberg_tables.BRONZE.iceberg_id)
+
         logging.info(f"dimensions: {dimensions}")
         for dimension_name, codes in dimensions.items():
             logging.info(f"dimension_name: {dimension_name}")
             logging.info(f"codes: {codes}")
             if len(codes) != 0:
-                not_in_codes = ",".join([f"'{code}'" for code in codes])
-                delete_from_branch_query = f"DELETE FROM {self.iceberg_tables.BRONZE.iceberg_id}.`branch_diss_tag_{self.tag_name}` WHERE {dimension_name} NOT IN ({not_in_codes})"
-                logging.info(f"delete_from_branch_query: {delete_from_branch_query}")
-                delete_from_branch_query_result = self.spark.sql(
-                    delete_from_branch_query
-                ).collect()
-
-                logging.info(
-                    f"result of delete_from_branch_query: {delete_from_branch_query_result}"
+                # not_in_codes = ",".join([f"'{code}'" for code in codes])
+                # delete_from_branch_query = f"DELETE FROM {self.iceberg_tables.BRONZE.iceberg_id}.`branch_diss_tag_{self.tag_name}` WHERE {dimension_name} NOT IN ({not_in_codes})"
+                # logging.info(f"delete_from_branch_query: {delete_from_branch_query}")
+                # delete_from_branch_query_result = self.spark.sql(
+                #     delete_from_branch_query
+                # ).collect()
+
+                # logging.info(
+                #     f"result of delete_from_branch_query: {delete_from_branch_query_result}"
+                # )
+                self.disseminated_tag_df = self.disseminated_tag_df.filter(
+                    col(dimension_name).isin(codes)
                 )
 
-        self.disseminated_tag_df = self.spark.read.option(
-            "branch", f"`diss_tag_{self.tag_name}`"
-        ).table(self.iceberg_tables.BRONZE.iceberg_id)
+        # self.disseminated_tag_df = self.spark.read.option(
+        #     "branch", f"`diss_tag_{self.tag_name}`"
+        # ).table(self.iceberg_tables.BRONZE.iceberg_id)
+
+        self.disseminated_tag_df.write.format("iceberg").mode("overwrite").save(
+            f"{self.iceberg_tables.BRONZE.iceberg_id}.`diss_tag_{self.tag_name}`"
+        )
 
         disseminated_tag_df = self.disseminated_tag_df.withColumn(
             "metadata", F.to_json(col("metadata"))

{sws_spark_dissemination_helper-0.0.93.dist-info → sws_spark_dissemination_helper-0.0.95.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sws-spark-dissemination-helper
-Version: 0.0.93
+Version: 0.0.95
 Summary: A Python helper package providing streamlined Spark functions for efficient data dissemination processes
 Project-URL: Repository, https://bitbucket.org/cioapps/sws-it-python-spark-dissemination-helper
 Author-email: Daniele Mansillo <danielemansillo@gmail.com>

{sws_spark_dissemination_helper-0.0.93.dist-info → sws_spark_dissemination_helper-0.0.95.dist-info}/RECORD

@@ -1,11 +1,11 @@
-sws_spark_dissemination_helper/SWSBronzeIcebergSparkHelper.py,sha256=ZPCpHgPVCsf7-7tWl6DDWgnXLkS02RoCvsomO3TmQ24,20418
+sws_spark_dissemination_helper/SWSBronzeIcebergSparkHelper.py,sha256=ja7zXIJNYdzvSiHdMB8oeKSQkLVu-FSUhy0vn_myhnU,21086
 sws_spark_dissemination_helper/SWSGoldIcebergSparkHelper.py,sha256=ZC7hxkppo6qmfCc2z5vm2Y2iH1901F-rx9Er9cxuzP4,16037
 sws_spark_dissemination_helper/SWSPostgresSparkReader.py,sha256=ja7AbOfbmC_EXHCJk7UMDzzbA-LRxzPkaaUmuvcihJ8,17449
 sws_spark_dissemination_helper/SWSSilverIcebergSparkHelper.py,sha256=zEppNq5shiHZH2yt5faWGsb5QEmpAQS0ToIrG6fmv6o,22231
 sws_spark_dissemination_helper/__init__.py,sha256=Efjoe9V4vGXWVp-DY5P6NbRwIUr_zkZJkDmMi-lf5Bc,262
 sws_spark_dissemination_helper/constants.py,sha256=hpHHlbojShMWRfyIelXz6c5BqFzO48Oap1zmztlMMrs,11349
 sws_spark_dissemination_helper/utils.py,sha256=6SzrXX0xhvynRyv-vRFDbc6V4UNe_RzKKETZAtefnhg,21341
-sws_spark_dissemination_helper-0.0.93.dist-info/METADATA,sha256=y1PL3ZygwfoyBxglsrNeP6IZvaUGTYCM03RuIjrqDMc,2823
-sws_spark_dissemination_helper-0.0.93.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-sws_spark_dissemination_helper-0.0.93.dist-info/licenses/LICENSE,sha256=zFzeb_j_6pXEHwH8Z0OpIkKFJk7vmhZjdem-K0d4zU4,1073
-sws_spark_dissemination_helper-0.0.93.dist-info/RECORD,,
+sws_spark_dissemination_helper-0.0.95.dist-info/METADATA,sha256=5oamHa9qI0T6_rNyZc4NLUM4kPUaCGfmHOypaNyky8s,2823
+sws_spark_dissemination_helper-0.0.95.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+sws_spark_dissemination_helper-0.0.95.dist-info/licenses/LICENSE,sha256=zFzeb_j_6pXEHwH8Z0OpIkKFJk7vmhZjdem-K0d4zU4,1073
+sws_spark_dissemination_helper-0.0.95.dist-info/RECORD,,