技術(shù)面試-oracle分析函數(shù)

createtableearnings--打工賺錢表(earnmonthvarchar2(6),--打工月份areavarchar2(20),--打工地區(qū)snovarchar2(10),--打工者編號(hào)snamevarchar2(20),--打工者timesint,--本月打工次數(shù)enumber(10,2),--每次賺enumber(10,2)--當(dāng)月總收入insertintoearningsvalues('200912','北平','511601','',11,30,11*30);insertintoearningsvalues('200912','北平','511602','大凱',8,25,8*25);insertintoearningsvalues('200912','北平','511603','',30,6.25,30*6.25);insertintoearningsvalues('200912','北平','511604','大亮',16,8.25,16*8.25);insertintoearningsvalues('200912','北平','511605','賤敬',30,11,30*11);insertintoearningsvalues('200912','金陵','511301','',15,12.25,15*12.25);insertintoearningsvalues('200912','金陵','511302','',27,16.67,27*16.67);insertintoearningsvalues('200912','金陵','511303','小妮',7,33.33,7*33.33);insertintoearningsvalues('200912','金陵','511304','小俐',0,18,0);insertintoearningsvalues('200912','金陵','511305','雪兒insertintoearningsvalues('201001','北平','511601','',0,30,0);insertintoearningsvalues('201001','北平','511602','大凱',14,25,14*25);insertintoearningsvalues('201001','北平','511603','',19,6.25,19*6.25);insertintoearningsvalues('201001','北平','511604','大亮',7,8.25,7*8.25);insertintoearningsvalues('201001','北平','511605','賤敬',21,11,21*11);insertintoearningsvalues('201001','金陵','511301','',6,12.25,6*12.25);insertintoearningsvalues('201001','金陵','511302','',17,16.67,17*16.67);insertintoearningsvalues('201001','金陵','511303','小妮',27,33.33,27*33.33);insertintoearningsvalues('201001','金陵','511304','小俐',16,18,16*18);insertintoearningsvalues('201001','金陵','511305','雪兒',11,9.88,11*9.88);select*fromsumselectearnmonth,area,sum( fromearningsgroupbyearnmonth,areaorderbyrollupselectearnmonth,area,sum( fromearningsgroupbyrollup(earnmonth,cube按照月份，地區(qū)進(jìn)行收入selectearnmonth,area,sum( fromearningsgroupbycube(earnmonth,area)orderbyearnmonth,areanullslast;groupbyearnmontharea以上三例都是先按照earnmonth分組，在earnmonth內(nèi)部再按area分組，并在area組內(nèi)統(tǒng) e總合groupbygroupgroupbyrollupgroupbyearnmonthgroupbycubeearnmontharea另外那個(gè)nullslast是把空值放在最后。(4)grou函在以上例子中，是用rollup和cube函數(shù)都會(huì)對(duì)結(jié)果集產(chǎn)生null，這時(shí)候可用grou函數(shù)來(lái)確認(rèn) 函數(shù)用法，帶一個(gè)參數(shù)，參數(shù)為字段名，1，Sql代 selectdecode(grou(earnmonth),1,'所有月份',earnmonth)月份,decode(grou(area),1,'全部地區(qū)',area)地區(qū), e)總金額fromgroupbycube(earnmonth,area)orderbyearnmonth,areanullslast;rank()overSql代 selectearnmonth月份area地區(qū),sname打工者,e收入rank()over(partitionbyearnmonth,areaorderby edesc)fromearnings;dense_rank()overSql代碼 selectarea,sname,edense_rank()over(partitionbyearnmonth,areaorderby edesc)paimingfromearnings;row_number()overSql代碼 selectearnmontharea,sname,erow_number()over(partitionbyearnmonth,areaorderby edesc)fromearnings;通過(guò)(5)(6)(7)rank,dense_rank,row_number結(jié)果集中如果出現(xiàn)兩個(gè)相同的數(shù)據(jù)，那么rank會(huì)進(jìn)行跳躍式的，但是dense_rank不會(huì)跳躍式的，兩個(gè)第二接下來(lái)還是第三row_number最牛，即使兩個(gè)數(shù)據(jù)相同，也不一樣sumSql代 selectearnmonth月份area地區(qū),sname打工者, e)over(partitionbyearnmonth,areaorder e)總收入fromSql代 selectdistinctearnmontharea e)over(partitionbyearnmonth,area) e)over(partitionbyearnmonth,area) e)over(partitionbyearnmonth,area) e)over(partitionbyearnmonth,fromlaglead求出每個(gè)打工者上個(gè)月和下個(gè)月有沒(méi)有賺錢 e大于零即為賺錢Sql代 selectearnmonth本月sname打工者 e,0),0,'沒(méi)賺','賺了'),1,0)over(partitionbysnameorderbyearnmonth)上月 e,0),0,'沒(méi)賺','賺了'),1,0)over(partitionbysnameorderbyearnmonth)下月from說(shuō)明：LagLeadNN（可以是其他字段的數(shù)據(jù)，比如根據(jù)字段甲查詢上lag(value_expression[,offset][,default])over([query_partition_clase]order_by_clause)；lead(value_expression[,offset][,default])over([query_partition_clase]order_by_clause)；value_expr