WITH AS短语,也叫做子查询部分(subquery factoring),可以做很多事情,定义一个SQL片断,该SQL片断会被整个SQL语句所用到。有的时候,是为了让SQL语句的可读性更高些,也有可能是在UNION ALL的不同部分,作为提供数据的部分。
特别对于UNION ALL比较有用。因为UNION ALL的每个部分可能相同,但是如果每个部分都去执行一遍的话,则成本太高,所以可以使用WITH AS短语,则只要执行一遍即可。如果WITH AS短语所定义的表名被调用两次以上,则优化器会自动将WITH AS短语所获取的数据放入一个TEMP表里,如果只是被调用一次,则不会。而提示materialize则是强制将WITH AS短语里的数据放入一个全局临时表里。
单个语法:
with tempName as (select ....) select ...
多个语法:
with tempName1 as (select ....), tempName2 as (select ....), tempName3 as (select ....) ... select ...
With查询语句不是以select开始的,而是以“WITH”关键字开头
可认为在真正进行查询之前预先构造了一个临时表TT,之后便可多次使用它做进一步的分析和处理。
例:从1-19中得到11-14。一般的sql如下:
select * from ( --模拟生一个20行的数据 SELECT LEVEL AS lv FROM DUAL CONNECT BY LEVEL < 20 ) tt WHERE tt.lv > 10 AND tt.lv < 15
使用With as 的SQL为:
with TT as( SELECT LEVEL AS lv FROM DUAL CONNECT BY LEVEL < 20 ) select lv from TT WHERE lv > 10 AND lv < 15;
多个临时表实例:
WITH T3 AS (SELECT T1.ID, T1.CODE1, T2.DESCRIPTION FROM TB_DATA T1, TB_CODE T2 WHERE T1.CODE1 = T2.CODE ), T4 AS (SELECT T1.ID, T1.CODE2, T2.DESCRIPTION FROM TB_DATA T1, TB_CODE T2 WHERE T1.CODE2 = T2.CODE ) SELECT T3.ID, T3.DESCRIPTION, T4.DESCRIPTION FROM T3, T4 WHERE T3.ID = T4.ID ORDER BY ID;
增加了SQL的易读性,如果构造了多个子查询,结构会更清晰;更重要的是:“一次分析,多次使用”,这也是为什么会提供性能的地方,达到了“少读”的目标。
上面第一种使用子查询的方法表被扫描了两次,而使用WITH Clause方法,表仅被扫描一次。这样可以大大的提高数据分析和查询的效率。
另外,观察WITH Clause方法执行计划,其中“SYS_TEMP_XXXX”便是在运行过程中构造的中间统计结果临时表。
4.1与select查询语句结合使用
查询同一个单据编号对应的借款单和核销单中,借款金额不相等的单据
with verificationInfo as (select ment.fnumber from shenzhenjm.t_finance_expenseremburseitem t left join shenzhenjm.t_finance_expenserembursement ment on ment.fid = t.fkrembursementid where 1 = 1 group by ment.fnumber), loanInfo as (select ment.fnumber from shenzhenjm.t_finance_expenseremburseitem2 t left join shenzhenjm.t_finance_expenserembursement ment on ment.fid = t.fkrembursementid where 1 = 1 group by ment.fnumber) select * from verificationInfo v, loanInfo l where l.fnumber = v.fnumber;
4.2与insert结合使用
如下的with as语句,不能放在insert前,而是放在紧接着要调用的地方前
insert into T_finance_ExpenseRemburseItem2 (FID, FKREMBURSEMENTID, FAMOUNT, FKCREATEBYID, FCREATETIME, FKCUID, FKCOSTTYPEID, FCOSTTYPENAME) with verificationInfo as (select ment.fnumber, sum(t.famount) vLoanSum, ment.fnumber "单据编号", sum(t.famount) "核销单中借款总额" from shenzhenjm.t_finance_expenseremburseitem t left join shenzhenjm.t_finance_expenserembursement ment on ment.fid = t.fkrembursementid where 1 = 1 group by ment.fnumber), loanInfo as (select ment.fnumber, sum(t.famount) loanSum, ment.fnumber "单据编号", sum(t.famount) "借款单中借款总额" from shenzhenjm.t_finance_expenseremburseitem2 t left join shenzhenjm.t_finance_expenserembursement ment on ment.fid = t.fkrembursementid where 1 = 1 group by ment.fnumber) select sys_guid(), ment.fid, t.famount, ment.fkcreatebyid, ment.fcreatetime, ment.fkcuid, t.fkcosttypeid, t.fcosttypename from T_finance_ExpenseRemburseItem t left join t_finance_expenserembursement ment on ment.fid = t.fkrembursementid where 1 = 1 and exists (select * from verificationInfo v, loanInfo l where l.fnumber = v.fnumber and l.loanSum != v.vLoanSum and v.fnumber = ment.fnumber);
4.3 与delete删除结合使用
delete from t_finance_expenseremburseitem2 item2 where exists(with temp as (select t.fnumber, sum(item1.famount) vloanSum, sum(item1.frealityamount) vSum, sum(item2.famount) loanSum from t_finance_expenserembursement t left join t_finance_expenseremburseitem item1 on item1.fkrembursementid = t.fid left join t_finance_expenseremburseitem2 item2 on item2.fkrembursementid = t.fid where 1 = 1 and t.frembursementtype = 'LOAN_REPORT' and to_char(t.fcreatetime, 'yyyy') > '2017' group by t.fnumber order by t.fnumber asc) select 1 from temp t left join t_finance_expenserembursement ment on t.fnumber = ment.fnumber left join t_finance_expenseremburseitem2 item on item.fkrembursementid = ment.fid where t.vloanSum != t.loanSum and item.fid = item2.fid);
4.4与update结合使用
update dest b set b.NAME = (with t as (select * from temp) select a.NAME from temp a where a.ID = b.ID)
其实就是把一大堆重复用到的SQL语句放在with as 里面,取一个别名,后面的查询就可以用它,这样对于大批量的SQL语句起到一个优化的作用,而且清楚明了。后面会分享更多devops和DBA方面的内容,感兴趣的朋友可以关注一下