工作中時長會用到豎錶轉橫表（列轉行）例如某商場每天都有營業額，資料庫中營業額儲存的方式是每天很多比每筆對應不同的消費記錄 ，可能有一天的營業額很多 對應的營業額明細就會很多，如果有個需求是要統計每天營業額多少，按天顯示結果該怎麼寫？ 下圖的例子是我們資料庫中一張記錄每月營業情況的表，可以看到201501有很多筆記錄，其他月份也有很多。 select amt_fee,inst_datefromR_KC_AGENT_STATISTICS wherei nst_date between 201501 and 201709; 實現方法很簡單，此處需要使用oracle的decode函式。如果使用的是mysql則需要使用case when。當讓oralce也可以使用case when來實現。

select sum(decode(inst_date, 201501, amt_fee, 0)) amtfees_201501,
       sum(decode(inst_date, 201502, amt_fee, 0)) amtfees_201502,
       sum(decode(inst_date, 201503, amt_fee, 0)) amtfees_201503,
       sum(decode(inst_date, 201504, amt_fee, 0)) amtfees_201504,
       sum(decode(inst_date, 201505, amt_fee, 0)) amtfees_201505,
       sum(decode(inst_date, 201506, amt_fee, 0)) amtfees_201506,
       sum(decode(inst_date, 201507, amt_fee, 0)) amtfees_201507,
       sum(decode(inst_date, 201508, amt_fee, 0)) amtfees_201508,
       sum(decode(inst_date, 201509, amt_fee, 0)) amtfees_201509,
       sum(decode(inst_date, 2015010, amt_fee, 0)) amtfees_201510,
       sum(decode(inst_date, 2015011, amt_fee, 0)) amtfees_201511,
       sum(decode(inst_date, 2015012, amt_fee, 0)) amtfees_201512
  from R_KC_AGENT_STATISTICS
 where inst_date between 201501 and 201512;
 

查詢結果如下： 如果有其他條件也可以在decode中再套一層decode。 再舉一個複雜點的例子 轉之前，同一個chnl_mcht_id,bank_mcht_id，app_key下 可能會有多個不同的acquire_type

select chnl_mcht_id,
       BANK_MCHT_ID,
       decode(max(wechats), 1, 'wechat', '') || '-' ||
       decode(max(alipays), 1, 'alipay', '') || '-' ||
       decode(max(qqs), 1, 'qq', '') || '-' ||
       decode(max(jds), 1, 'jd', '') || '-' ||
       decode(max(baidus), 1, 'baidus', ''),
       APP_KEY
  from (select chnl_mcht_id,
               BANK_MCHT_ID,
               decode(ACQUIRER_TYPE, 'wechat', 1, 0) wechats,
               decode(ACQUIRER_TYPE, 'alipay', 1, 0) alipays,
               decode(ACQUIRER_TYPE, 'qq', 1, 0) qqs,
               decode(ACQUIRER_TYPE, 'jd', 1, 0) jds,
               decode(ACQUIRER_TYPE, 'baidu', 1, 0) baidus,
               APP_KEY
          from (SELECT DISTINCT chnl_mcht_id,
                                BANK_MCHT_ID,
                                ACQUIRER_TYPE,
                                APP_KEY
                  FROM wallet_mcht_info_reg
                 WHERE chnl_mcht_id IS NOT NULL
                 group by chnl_mcht_id, BANK_MCHT_ID, APP_KEY, ACQUIRER_TYPE)
         group by chnl_mcht_id, BANK_MCHT_ID, APP_KEY, ACQUIRER_TYPE)
 group by chnl_mcht_id, BANK_MCHT_ID, APP_KEY
 

轉之後同一個chnl_mcht_id,bank_mcht_id，app_key下面 不同的acquire_type 被壓縮排一行