sql - 列更新 Oracle SQL

Question

我有以下格式的数据:

Tran No   Date        Store ID   Register ID  Tender
  01      23-10-2015    1000        001        CASH
  01      23-10-2015    1000        001        CRDT
  02      23-10-2015    1000        001        CASH
  02      23-10-2015    1000        001        GIFT

表中已添加新列 (Tran Seq)

这样新数据就会变成这样

Tran No   Date        Store ID   Register ID  Tender   Tran Seq   
  01      23-10-2015    1000        001        CASH       0
  01      23-10-2015    1000        001        CRDT       1
  02      23-10-2015    1000        001        CASH       0
  02      23-10-2015    1000        001        GIFT       1

对于每个相同的 TRAN NO、DATE、STORE ID 和 REGISTER ID，每个行项目都会在 Tran Seq 字段中获得一个新的序列号。

如何实现上述结果？

Answer 1

您可以使用ROW_NUMBER()分析函数:

例如，

设置

CREATE TABLE t
    (tran_no NUMBER, 
    "DATE" VARCHAR2(10),
     store_id NUMBER, 
     Register_ID NUMBER, 
     Tender varchar2(4));

INSERT ALL 
    INTO t (tran_no, "DATE", store_id, Register_ID, Tender)
         VALUES (01, '23-10-2015', 1000, 001, 'CASH')
    INTO t (tran_no, "DATE", store_id, Register_ID, Tender)
         VALUES (01, '23-10-2015', 1000, 001, 'CRDT')
    INTO t (tran_no, "DATE", store_id, Register_ID, Tender)
         VALUES (02, '23-10-2015', 1000, 001, 'CASH')
    INTO t (tran_no, "DATE", store_id, Register_ID, Tender)
         VALUES (02, '23-10-2015', 1000, 001, 'GIFT')
    INTO t (tran_no, "DATE", store_id, Register_ID, Tender)
         VALUES (02, '23-10-2015', 1000, 001, 'CRDT')
SELECT * FROM dual;

查询

SQL> SELECT t.*,
  2    row_number() OVER(PARTITION BY tran_no,
  3                                   "DATE",
  4                                   store_id,
  5                                   register_id
  6                      ORDER BY tender) - 1 tran_seq
  7  FROM t;

   TRAN_NO DATE         STORE_ID REGISTER_ID TEND   TRAN_SEQ
---------- ---------- ---------- ----------- ---- ----------
         1 23-10-2015       1000           1 CASH          0
         1 23-10-2015       1000           1 CRDT          1
         2 23-10-2015       1000           1 CASH          0
         2 23-10-2015       1000           1 CRDT          1
         2 23-10-2015       1000           1 GIFT          2

SQL>

注意:

1. 列名称之间不能有空格。我用下划线代替。

2. 您不能使用 Oracle 关键字。如果要使用，那么就需要使用双引号。例如，“日期”列。请记住，您需要在引用的所有地方使用双引号。

来自 Database Object Naming Rules 的文档:

A quoted identifier begins and ends with double quotation marks ("). If you name a schema object using a quoted identifier, then you must use the double quotation marks whenever you refer to that object.

更新

OP 希望使用生成的序列更新新列。您可以使用MERGE语句。

SQL> ALTER TABLE t ADD tran_seq NUMBER;

Table altered.

SQL> MERGE INTO t
  2  USING(
  3  SELECT TRAN_NO, "DATE", STORE_ID, REGISTER_ID, TENDER,
  4    row_number() OVER(PARTITION BY tran_no,
  5                                   "DATE",
  6                                   store_id,
  7                                   register_id
  8                      ORDER BY tender) - 1 tran_seq
  9  FROM t
 10  ) s
 11  ON (t.tran_no = s.tran_no
 12      AND t."DATE" = s."DATE"
 13      AND t.store_id = s.store_id
 14      AND t.register_id = s.register_id
 15      AND t.tender = s.tender
 16      )
 17  WHEN MATCHED THEN
 18     UPDATE SET t.tran_seq = s.tran_seq
 19  /

5 rows merged.

SQL> SELECT * FROM t ORDER BY tran_no, tran_seq;

   TRAN_NO DATE         STORE_ID REGISTER_ID TEND   TRAN_SEQ
---------- ---------- ---------- ----------- ---- ----------
         1 23-10-2015       1000           1 CASH          0
         1 23-10-2015       1000           1 CRDT          1
         2 23-10-2015       1000           1 CASH          0
         2 23-10-2015       1000           1 CRDT          1
         2 23-10-2015       1000           1 GIFT          2