一，列表分組之後查詢所有組內某一個時間欄位最大的資料

ContractEndDate ：某一個時間欄位
RECORDID：表ID

SELECT A.* from (select * from (SELECT * FROM hrm_contract order by ContractEndDate DESC limit 10000) a GROUP BY a.A00) B
 left join hrm_contract A on a.RECORDID=b.RECORDID

二，多級選單表根據parent_id查詢包括父節點在內的所有子節點id

SELECT * FROM (

SELECT t1.id,IF(FIND_IN_SET(t1.parent_id,@pids)>0,@pids:=CONCAT(@pids,',',id),0) as ischild FROM (
SELECT * FROM menu  ORDER BY  parent_id
) t1,(SELECT @pids:=1) t2

UNION
SELECT dd.id ,dd.id as ischild FROM menu as dd WHERE dd.id=1
) cc WHERE cc.ischild <> 0
 

三，多級選單表根據子節點id查詢包括子節點在內的所有父節點parent_id

表結構

查詢sql

SELECT T2.id
FROM ( 
    SELECT 
        @r AS _id, 
        (SELECT @r := parent_id FROM user WHERE id = _id) AS parent_id, 
        @l := @l + 1 AS lvl 
    FROM 
        (SELECT @r := 7, @l := 0) vars,  -- 查詢id為5的所有上級
        user h 
    WHERE @r <> 0) T1 
JOIN user T2 
ON T1._id = T2.id 
ORDER BY T1.lvl DESC
 

查詢結果

四，group by 對多個欄位進行分組

group by 欄位一：將所有具有相同欄位一的記錄放到一個分組裡

group by 欄位一，欄位二 ：將所有具有相同欄位一和欄位二的記錄放到一個分組裡

五，Case When和聚合函式count、sum的使用

根據貧困標誌，取出不同標準下的的人數有多少，以區縣位單位。

1代表已脫貧人口；2代表返貧人口；3代表未脫貧

select 
t3.region_id as 'regionId',
t3.region_name as 'regionName',
count(case when t1.tricolor = 1 then 1 else null end ) as 'greenTotal',
count(case when t1.tricolor = 2 then 1 else null end ) as 'yellowTotal',
count(case when t1.tricolor = 3 then 1 else null end )as 'redTotal'
from pa_household_member t1
left join pa_household t2 on t2.household_id = t1.household_id
left join system_region t3 on t3.region_id = t2.area 
group by t2.area
 

count(1) 算上一條記錄 count（null）為0；利用此特性進行分類統計。不用寫複雜的子查詢。count在套住case when ..... 例如： count (case when ....end )