no message

Pyspark_job/sqoop_export/dwt_aba_last365.py

@@ -37,12 +37,15 @@ if __name__ == '__main__':
     with engine.connect() as connection:
         sql = f"""
             drop table if exists {export_tb_rel};
-            create table if not exists {export_tb_rel} 
+            create table if not exists {export_tb_rel}
             (
                 like {export_tb_before} including comments
             );
             ALTER TABLE {export_tb_rel} ALTER COLUMN st_movie_brand_label TYPE VARCHAR(20);
             ALTER TABLE {export_tb_rel} ALTER COLUMN total_appear_month TYPE VARCHAR(50);
+            -- st_attribute_label 在正式表是 VARCHAR[]，Sqoop 无法直接写入数组类型
+            -- 先临时改成 VARCHAR(200) 让 Sqoop 写字符串，交换完成后再 ALTER 回 VARCHAR[]
+            ALTER TABLE {export_tb_rel} ALTER COLUMN st_attribute_label TYPE VARCHAR(200);
         """
         print("================================执行sql================================")
         print(sql)
@@ -187,7 +190,9 @@ if __name__ == '__main__':
             "market_cycle_type",
             "rank_lastest",
             "rank_change_rate_lastest",
-            "rank_rate_of_change_lastest"
+            "rank_rate_of_change_lastest",
+            # 搜索词属性标签(12 个月合并去重,逗号分隔字符串,sqoop 导出 PG 后转 VARCHAR[])
+            "st_attribute_label"
         ],
         partition_dict={
             "site_name": site_name,
@@ -216,6 +221,13 @@ if __name__ == '__main__':
             ALTER COLUMN total_appear_month TYPE INTEGER[]
             USING string_to_array(total_appear_month, ',')::int[];
 
+            -- 交换完成后，把 st_attribute_label 从 VARCHAR 转回 VARCHAR[]
+            -- 用 string_to_array 把 Sqoop 写入的逗号串（如 "材质,颜色"）拆成数组（如 {{材质,颜色}}）
+            -- 词典无匹配的词在 PySpark 端已 fillna "-1"，转换后是 {{-1}}，与 Java 占位约定一致
+            ALTER TABLE {export_tb_before}
+            ALTER COLUMN st_attribute_label TYPE VARCHAR[]
+            USING string_to_array(st_attribute_label, ',')::varchar[];
+
             alter table {export_tb_before} drop if exists keyword_tsv;
             alter table {export_tb_before} add column keyword_tsv tsvector generated always as (to_tsvector('english_amazonword', search_term)) STORED;
             drop index if exists {export_tb_before}_keyword_tsv_idx;