标签：
准备数据
浙江,杭州,300
浙江,宁波,150
浙江,温州,200
浙江,嘉兴,100
江苏,南京,270
江苏,苏州,299
江苏,某市,200
江苏,某某市,100
?
创建表
CREATE table pcp
(province string,city string,people int)
ROW FORMAT DELIMITED FIELDS TERMINATED BY ‘,‘
STORED AS TEXTFILE;
?
导入数据
load data inpath ‘/tmp/1.txt‘ into table pcp;
?
普通查询
select * from pcp order by people desc;
浙江????杭州????300
浙江????宁波????150
浙江????温州????200
浙江????嘉兴????100
江苏????南京????270
江苏????苏州????299
江苏????某市????200
江苏????某某市????100
?
?
综合查询
select province,city,
rank() over (order by people desc) rank,
dense_rank() over (order by people desc) dense_rank,
row_number() over(order by people desc) row_number
from pcp
group by province,city,people;
?
浙江????杭州????300????1????1????1
江苏????苏州????299????2????2????2
江苏????南京????270????3????3????3
江苏????某市????200????4????4????4
浙江????温州????200????4????4????5
浙江????宁波????150????6????5????6
江苏????某某市????100????7????6????7
浙江????嘉兴????100????7????6????8
?
主要注意打圈的:
?row_number顺序下来,
rank:在遇到数据相同项时,会留下空位,(红框内第一列,4,4,6)
dense_rank:在遇到数据相同项时,不会留下空位,(红框内第一列,4,4,5)
?
分组统计查询
select province,city,
rank() over (partition by province order by people desc) rank,
dense_rank() over (partition by province order by people desc) dense_rank,
row_number() over(partition by province order by people desc) row_number
from pcp
group by province,city,people;
江苏????苏州????299????1????1????1
江苏????南京????270????2????2????2
江苏????某市????200????3????3????3
江苏????某某市????100????4????4????4
浙江????杭州????300????1????1????1
浙江????温州????200????2????2????2
浙江????宁波????150????3????3????3
浙江????嘉兴????100????4????4????4
?
Hive ROW_NUMBER,RANK(),DENSE_RANK()
标签：
原文地址：http://www.cnblogs.com/dycg/p/4260283.html
