Merge branch 'developer' of http://47.106.101.75/abel_cjy/Amazon-Selection-Data into developer

Pyspark_job/img_search/img_local_path.py

@@ -38,7 +38,7 @@ class AsinImageLocalPath(object):
     def read_data(self):
         # sql = f"select img_unique from selection.{self.image_table} where state=1;"
         # df = pd.read_sql(sql, con=self.engine_doris)
-        sql = f"SELECT * from us_inv_img_info WHERE updated_at>= CURDATE() - INTERVAL 3 DAY and state =3;"
+        sql = f"SELECT * from us_inv_img_info WHERE updated_at>= CURDATE() - INTERVAL 3 DAY and state =3 and id_segment is not null;"
         df = pd.read_sql(sql, con=self.engine_mysql)
         print(f"sql: {sql}", df.shape)
         return df

Pyspark_job/sqoop_export/dwd_buyer_st_pg.py

@@ -61,14 +61,14 @@ if __name__ == '__main__':
     if date_type == 'us':
         sql = f"""
             create table if not exists {export_tb} (
-                like {export_master_tb} including comments
+                like {export_master_tb} including defaults including comments
             );
             delete from {export_tb} where report_date >= '{start_day}' and report_date <= '{end_day}'
         """
     else:
         sql = f"""
                   create table if not exists {export_tb} (
-                      like {export_master_tb} including comments
+                      like {export_master_tb} including  defaults including comments
                   );
                   delete from {export_tb} where report_date > '{start_day}' and report_date <= '{end_day}'
               """