常規情況下的sql分組統計為:

 select count(1) from 表 where 條件 group by 欄位;

但是有時往往需要新增不同的條件已經去重的統計以上語句就不能滿足需求。

解決方案為：

1.新增條件的統計方案：

COUNT(CASE WHEN 條件 THEN 1 ELSE NULL END) xxx GROUP BY 分組欄位

2.新增條件並去重的統計方案：

COUNT(DISTINCT CASE WHEN 條件 THEN 去重欄位 END) xxx GROUP BY 分組欄位

綜合示例：

SELECT
 
 dc.user_sources AS sources,
COUNT(CASE WHEN dc.`count_type` IN (1,4) THEN 1 ELSE NULL END) AS djNum1,
COUNT(CASE WHEN dc.`count_type` IN (2,5) THEN 1 ELSE NULL END) AS djNum2,
COUNT(CASE WHEN dc.`count_type` IN (3,6) THEN 1 ELSE NULL END) AS djNum3,
COUNT(DISTINCT CASE WHEN dc.`count_type` IN (1,4) THEN
 
 dc.`user_id` END) AS fwNum1,
COUNT(DISTINCT CASE WHEN dc.`count_type` IN (2,5) THEN dc.`user_id` END) AS fwNum2,
COUNT(DISTINCT CASE WHEN dc.`count_type` IN (3,6) THEN dc.`user_id` END) AS fwNum3,
COUNT(DISTINCT CASE WHEN dc.`count_type` IN (2,5) THEN dc.`user_id` END) AS fwNumc4,
COUNT(DISTINCT CASE
 
 WHEN dc.`count_type` IN (3,6) THEN dc.`user_id` END) AS fwNumc5
FROM `credit_dc_project_count` dc WHERE 1=1
AND dc.user_sources IN('wodong' , 'qq' , 'ydb_dkw' , 'chh_12d' , '12d' , 'jd_dkw' , 'hds_dkw' , 'ksd_12d' , 'ttym_dkw' , 'ios' , 'dkwaaa' , 'gzh' , 'chaomi' , 'mmd_12d' , 'ydb_12d' , 'hjsd_dkw' , 'papadai' , 'chd_dkw')
GROUP BY dc.user_sources