Transpose String colomn values to Row Oracle
我的初始表是这样的。
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 | CREATE TABLE pivot_string ( col1 NUMBER, col2 NUMBER, col3 VARCHAR2(6), col3_id NUMBER, col3_desc VARCHAR2(10) ) ; INSERT INTO pivot_string SELECT 123, 9875,'RO', 40, 'Roma' FROM dual; INSERT INTO pivot_string SELECT 123, 9875,'IT', 40, 'iteration' FROM dual; INSERT INTO pivot_string SELECT 123, 9875,'US', 78, 'world' FROM dual; INSERT INTO pivot_string SELECT 123, 9875,'WE', 56, 'WHAT' FROM dual; COMMIT; SELECT * FROM pivot_string; -- INSERT INTO pivot_string SELECT 123, 4875,'RO', 55, 'Roma' FROM dual; --INSERT INTO pivot_string SELECT 124, 4875,'RO', 44, 'Roma' FROM dual; TABLE: pivot_string' ------------------------------------------ col1 | col2 | col3 | col3_id | col3_desc ----------------------------------------- 123 | 9875 | 'RO' | 40 | 'Roma' 123 | 9875 | 'IT' | 50 | 'iteration' 123 | 9875 | 'US' | 78 | 'world' 123 | 9875 | 'WE' | 56 | 'WHAT' Excepted- pivot_string -------------------------------------------------------------------------- col1 | col2 | ro | ro_desc | it | it_desc | us | us_desc | we | we_desc --------------------------------------------------------------------------- 123 | 9875 | 40 | 'Roma' | 50 | 'iteration' | 78 | 'world' | 56 | 'WHAT' |
我的方法
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 | WITH mytest AS( SELECT a.col1, a.col2, a.col3,a.col3_id, a.col3_desc FROM pivot_string a ) SELECT COALESCE(RO.col1,IT.col1, US.col1, WE.col1)col1, COALESCE(RO.col2,IT.col2, US.col2, WE.col2) col2, RO,RO_desc IT,IT_desc, US,US_desc, WE, WE_desc FROM( (SELECT col1,col2, col3_id RO, col3_desc RO_desc FROM mytest WHERE col3 = 'RO' )RO LEFT JOIN (SELECT col1,col2, col3_id IT, col3_desc IT_desc FROM mytest WHERE col3 = 'IT' ) IT ON(RO.col1 =IT.col1 AND RO.col2 = IT.col2) LEFT JOIN ( SELECT col1,col2, col3_id US, col3_desc US_desc FROM mytest WHERE col3 = 'US' )US ON(RO.col1 =US.col1 AND RO.col2 = US.col2) LEFT JOIN (SELECT col1,col2, col3_id WE, col3_desc WE_desc FROM mytest WHERE col3 = 'WE') WE ON(RO.col1 =WE.col1 AND RO.col2 = WE.col2) ) ; |
我解决这个问题的方法不好,因为 col3 有超过 100 个不同的值。
所以像我一样左加入所有肯定不是正确的方法。
由于 col3 和 col3_desc,我不能使用 pivot。它们的类型是 Varchar2
有人知道解决此类问题的更好方法吗?
我前几天看到类似的问题。您可以在 PIVOT 中添加多个聚合函数。检查以下查询:
1 2 3 4 5 6 7 | SELECT * FROM ( SELECT * FROM pivot_string ) PIVOT ( SUM(COL3_ID) AS id,MIN(COL3_DESC) AS descp FOR col3 IN ('RO' AS RO, 'IT' AS IT, 'US' AS US, 'WE' AS WE) ); |
参考:
在 Oracle 行的多个列上使用数据透视
SQL FIDDLE