在GROUP BY級別使用「CASE」命令

Question

我試圖獲得5列，第一列是評估單元（EU），但在接下來的4列中，我要計算有多少家庭（不同的INSTANCE_ID_HOUSE）符合特定標準，但數據由居民儲存。這裏有4個查詢我使用的（我基本上只是想將它們組合成一個查詢）：

SELECT EU, COUNT(DISTINCT INSTANCE_ID_HOUSE) FROM SENEGAL_CLEAN 
WHERE CLUSTER_COMPLETE>0 
GROUP BY EU 
ORDER BY EU ASC 

SELECT COUNT(DISTINCT INSTANCE_ID_HOUSE) FROM SENEGAL_CLEAN 
WHERE CLUSTER_COMPLETE>0 AND (W1_DRINK_SOURCE=1 OR W1_DRINK_SOURCE=4) 
GROUP BY EU 
ORDER BY EU ASC 

SELECT COUNT(DISTINCT INSTANCE_ID_HOUSE) FROM SENEGAL_CLEAN 
WHERE CLUSTER_COMPLETE>0 AND (W2_GET_DRINK_WATER=1 OR W2_GET_DRINK_WATER=2) 
GROUP BY EU 
ORDER BY EU ASC 

SELECT COUNT(DISTINCT INSTANCE_ID_HOUSE) FROM SENEGAL_CLEAN 
WHERE CLUSTER_COMPLETE>0 AND (S2_SEE_LATRINE=1 OR S2_SEE_LATRINE=2) 
GROUP BY EU 
ORDER BY EU ASC 

Answer 1

可以使用case語句爲您計數：

select 
    eu, 
    count (distinct case when cluster_complete > 0 then instance_id_house else null end) as YourAliasHere, 
    count (distinct when cluster_complete > 0 and (W1_DRINK_SOURCE=1 OR W1_DRINK_SOURCE=4) then instance_id_house else null end) as YourNextAlias\ 
    ... 
FROM SENEGAL_CLEAN 
group by EU 

所以你指望每個instance_id_house它符合每個案例陳述中的標準。我認爲你可以留下你的where子句（根據你現在展示的內容）。

Answer 2

嘗試是這樣的：

Select S.EU,t1.c1 , t2.c2, t3.c3, t4.c4 
From SENEGAL_CLEAN S 
inner join(

SELECT EU, COUNT(DISTINCT INSTANCE_ID_HOUSE) c1 FROM SENEGAL_CLEAN 
WHERE CLUSTER_COMPLETE>0 
GROUP BY EU 
)t1 on t1.Eu= s.Eu 

inner join(
SELECT EU,COUNT(DISTINCT INSTANCE_ID_HOUSE) c2 FROM SENEGAL_CLEAN 
WHERE CLUSTER_COMPLETE>0 AND (W1_DRINK_SOURCE=1 OR W1_DRINK_SOURCE=4) 
GROUP BY EU 
)t2 on t2.Eu= s.Eu 

inner join(
SELECT Eu,COUNT(DISTINCT INSTANCE_ID_HOUSE) c3 FROM SENEGAL_CLEAN 
WHERE CLUSTER_COMPLETE>0 AND (W2_GET_DRINK_WATER=1 OR W2_GET_DRINK_WATER=2) 
GROUP BY EU 
)t3 on t3.Eu= s.Eu 

inner join(
SELECT Eu,COUNT(DISTINCT INSTANCE_ID_HOUSE) c4 FROM SENEGAL_CLEAN 
WHERE CLUSTER_COMPLETE>0 AND (S2_SEE_LATRINE=1 OR S2_SEE_LATRINE=2) 
GROUP BY EU 
)t4 on t4.Eu = s.Eu 

SqlFiddle：http://sqlfiddle.com/#!9/90b33/1