select to_char(Applydate, 'yyyy/mm' ) As Month,
count(Decode(FormID, 'A01', SerialID, null )) AS A01,
count(Decode(FormID, 'A02', SerialID, null )) AS A02
from AFS_FLOW group by to_char(Applydate, 'yyyy/mm' )
這個語法可以將資料表:
Applydate FormID SerialID
========= ====== ========
2007/09/11 A01 T0000001
2007/09/12 A02 T0000002
2007/10/11 A01 T0000003
2007/10/12 A01 T0000004
變成這樣:
Month A01 A02
========= ====== ======
2007/09 1 1
2007/10 2 0
這個語法有個缺點是你必須事先知道橫向欄位的名稱及數量,如果欄位是變數就無效了。
沒有留言:
張貼留言