将 PL/SQL 个程序转换为 Pyspark

Question

BEGIN
   open v_refcur for
       SELECT A.LGCY_LNDR_NO
            , A.LGCY_LNDR_BR_NO
            , A.LNDR_NM
            , B.ADDR_LINE1_TXT
            , B.ADDR_LINE2_TXT
            , B.CITY_NM
            , B.ST_CD
            , B.POSTAL_CD
            , C.FAX_NO
        FROM LNDR_CUST_XREF A
        LEFT OUTER JOIN LNDR_CUST_ADDR B
        ON A.LNDR_ID = B.LNDR_ID
        AND B.ADDR_TYP_CD = 'MAIL'
        LEFT OUTER JOIN LNDR_CUST_ADDR C
        ON A.LNDR_ID = C.LNDR_ID
        AND C.ADDR_TYP_CD = 'SITE'
        WHERE A.LGCY_LNDR_NO = LNDR_NO
        AND A.LGCY_LNDR_BR_NO = BRN_NO
        AND A.TA_CUST_FLG = 'Y';
   SQL_CD := SWV_SQLCODE;
END;

以上代码逐行转换会是什么？我手头没有数据库，那么 Pyspark 中 PL/SQL 代码最合适的要点是什么？

Answer 1

这个语句可以是re-written类似下面的东西-

df = (df_LNDR_CUST_XREF.alias('A')
      .df_LNDR_CUST_ADDR.alias('B'), ((A.LNDR_ID == B.LNDR_ID) & (B.ADDR_TYP_CD == 'MAIL'), "left")
      .df_LNDR_CUST_ADDR.alias('C'), ((A.LNDR_ID == C.LNDR_ID) & (C.ADDR_TYP_CD == 'SITE'), "left")
      .where ((A.LGCY_LNDR_NO == LNDR_NO) & (A.LGCY_LNDR_BR_NO == BRN_NO) & (A.TA_CUST_FLG == 'Y'))
      .select (F.col("A.LGCY_LNDR_NO")
            , F.col("A.LGCY_LNDR_BR_NO")
            , F.col("A.LNDR_NM")
            , F.col("B.ADDR_LINE1_TXT")
            , F.col("B.ADDR_LINE2_TXT")
            , F.col("B.CITY_NM")
            , F.col("B.ST_CD")
            , F.col("B.POSTAL_CD")
            , F.col("C.FAX_NO"))
     )

不过我还没有测试过。

将 PL/SQL 个程序转换为 Pyspark

Converting PL/SQL procedures to Pyspark

python

plsql

pyspark