Hive分析窗口函数(四) LAG,LEAD,FIRST_VALUE,LAST_VALUE

问题导读

1.LAG功能是什么?
2.LEAD与LAG功能有什么相似的地方那个?
3.FIRST_VALUE与LAST_VALUE分别完成什么功能?

 

继续学习这四个分析函数(LAG,LEAD,FIRST_VALUE,LAST_VALUE)。注意: 这几个函数不支持WINDOW子句。


Hive版本为 apache-hive-0.13.1

数据准备:

cookie1,2019-04-10 10:00:02,url2
cookie1,2019-04-10 10:00:00,url1
cookie1,2019-04-10 10:03:04,1url3
cookie1,2019-04-10 10:50:05,url6
cookie1,2019-04-10 11:00:00,url7
cookie1,2019-04-10 10:10:00,url4
cookie1,2019-04-10 10:50:01,url5
cookie2,2019-04-10 10:00:02,url22
cookie2,2019-04-10 10:00:00,url11
cookie2,2019-04-10 10:03:04,1url33
cookie2,2019-04-10 10:50:05,url66
cookie2,2019-04-10 11:00:00,url77
cookie2,2019-04-10 10:10:00,url44
cookie2,2019-04-10 10:50:01,url55

CREATE EXTERNAL TABLE zcg1234 (
cookieid string,
createtime string,  --页面访问时间
url STRING       --被访问页面
) ROW FORMAT DELIMITED 
FIELDS TERMINATED BY ',' 
stored as textfile location '/tmp/zcg1234/';


hive> select * from zcg1234;
OK
cookie1 2019-04-10 10:00:02     url2
cookie1 2019-04-10 10:00:00     url1
cookie1 2019-04-10 10:03:04     1url3
cookie1 2019-04-10 10:50:05     url6
cookie1 2019-04-10 11:00:00     url7
cookie1 2019-04-10 10:10:00     url4
cookie1 2019-04-10 10:50:01     url5
cookie2 2019-04-10 10:00:02     url22
cookie2 2019-04-10 10:00:00     url11
cookie2 2019-04-10 10:03:04     1url33
cookie2 2019-04-10 10:50:05     url66
cookie2 2019-04-10 11:00:00     url77
cookie2 2019-04-10 10:10:00     url44
cookie2 2019-04-10 10:50:01     url55

LAG
LAG(col,n,DEFAULT) 用于统计窗口内往上第n行值
第一个参数为列名,第二个参数为往上第n行(可选,默认为1),第三个参数为默认值(当往上第n行为NULL时候,取默认值,如不指定,则为NULL)

SELECT cookieid,
createtime,
url,
ROW_NUMBER() OVER(PARTITION BY cookieid ORDER BY createtime) AS rn,
LAG(createtime,1,'1970-01-01 00:00:00') OVER(PARTITION BY cookieid ORDER BY createtime) AS last_1_time,
LAG(createtime,2) OVER(PARTITION BY cookieid ORDER BY createtime) AS last_2_time 
FROM zcg1234;


cookieid createtime             url    rn       last_1_time             last_2_time
-------------------------------------------------------------------------------------------
cookie1 2019-04-10 10:00:00     url1    1       1970-01-01 00:00:00     NULL
cookie1 2019-04-10 10:00:02     url2    2       2019-04-10 10:00:00     NULL
cookie1 2019-04-10 10:03:04     1url3   3       2019-04-10 10:00:02     2019-04-10 10:00:00
cookie1 2019-04-10 10:10:00     url4    4       2019-04-10 10:03:04     2019-04-10 10:00:02
cookie1 2019-04-10 10:50:01     url5    5       2019-04-10 10:10:00     2019-04-10 10:03:04
cookie1 2019-04-10 10:50:05     url6    6       2019-04-10 10:50:01     2019-04-10 10:10:00
cookie1 2019-04-10 11:00:00     url7    7       2019-04-10 10:50:05     2019-04-10 10:50:01
cookie2 2019-04-10 10:00:00     url11   1       1970-01-01 00:00:00     NULL
cookie2 2019-04-10 10:00:02     url22   2       2019-04-10 10:00:00     NULL
cookie2 2019-04-10 10:03:04     1url33  3       2019-04-10 10:00:02     2019-04-10 10:00:00
cookie2 2019-04-10 10:10:00     url44   4       2019-04-10 10:03:04     2019-04-10 10:00:02
cookie2 2019-04-10 10:50:01     url55   5       2019-04-10 10:10:00     2019-04-10 10:03:04
cookie2 2019-04-10 10:50:05     url66   6       2019-04-10 10:50:01     2019-04-10 10:10:00
cookie2 2019-04-10 11:00:00     url77   7       2019-04-10 10:50:05     2019-04-10 10:50:01


last_1_time: 指定了往上第1行的值,default为'1970-01-01 00:00:00'  
             cookie1第一行,往上1行为NULL,因此取默认值 1970-01-01 00:00:00
             cookie1第三行,往上1行值为第二行值,2019-04-10 10:00:02
             cookie1第六行,往上1行值为第五行值,2019-04-10 10:50:01
last_2_time: 指定了往上第2行的值,为指定默认值
             cookie1第一行,往上2行为NULL
             cookie1第二行,往上2行为NULL
             cookie1第四行,往上2行为第二行值,2019-04-10 10:00:02
             cookie1第七行,往上2行为第五行值,2019-04-10 10:50:01

LEAD

与LAG相反
LEAD(col,n,DEFAULT) 用于统计窗口内往下第n行值
第一个参数为列名,第二个参数为往下第n行(可选,默认为1),第三个参数为默认值(当往下第n行为NULL时候,取默认值,如不指定,则为NULL)

SELECT cookieid,
createtime,
url,
ROW_NUMBER() OVER(PARTITION BY cookieid ORDER BY createtime) AS rn,
LEAD(createtime,1,'1970-01-01 00:00:00') OVER(PARTITION BY cookieid ORDER BY createtime) AS next_1_time,
LEAD(createtime,2) OVER(PARTITION BY cookieid ORDER BY createtime) AS next_2_time 
FROM zcg1234;


cookieid createtime             url    rn       next_1_time             next_2_time 
-------------------------------------------------------------------------------------------
cookie1 2019-04-10 10:00:00     url1    1       2019-04-10 10:00:02     2019-04-10 10:03:04
cookie1 2019-04-10 10:00:02     url2    2       2019-04-10 10:03:04     2019-04-10 10:10:00
cookie1 2019-04-10 10:03:04     1url3   3       2019-04-10 10:10:00     2019-04-10 10:50:01
cookie1 2019-04-10 10:10:00     url4    4       2019-04-10 10:50:01     2019-04-10 10:50:05
cookie1 2019-04-10 10:50:01     url5    5       2019-04-10 10:50:05     2019-04-10 11:00:00
cookie1 2019-04-10 10:50:05     url6    6       2019-04-10 11:00:00     NULL
cookie1 2019-04-10 11:00:00     url7    7       1970-01-01 00:00:00     NULL
cookie2 2019-04-10 10:00:00     url11   1       2019-04-10 10:00:02     2019-04-10 10:03:04
cookie2 2019-04-10 10:00:02     url22   2       2019-04-10 10:03:04     2019-04-10 10:10:00
cookie2 2019-04-10 10:03:04     1url33  3       2019-04-10 10:10:00     2019-04-10 10:50:01
cookie2 2019-04-10 10:10:00     url44   4       2019-04-10 10:50:01     2019-04-10 10:50:05
cookie2 2019-04-10 10:50:01     url55   5       2019-04-10 10:50:05     2019-04-10 11:00:00
cookie2 2019-04-10 10:50:05     url66   6       2019-04-10 11:00:00     NULL
cookie2 2019-04-10 11:00:00     url77   7       1970-01-01 00:00:00     NULL

--逻辑与LAG一样,只不过LAG是往上,LEAD是往下。

FIRST_VALUE

取分组内排序后,截止到当前行,第一个值

SELECT cookieid,
createtime,
url,
ROW_NUMBER() OVER(PARTITION BY cookieid ORDER BY createtime) AS rn,
FIRST_VALUE(url) OVER(PARTITION BY cookieid ORDER BY createtime) AS first1 
FROM zcg1234;

cookieid  createtime            url     rn      first1
---------------------------------------------------------
cookie1 2019-04-10 10:00:00     url1    1       url1
cookie1 2019-04-10 10:00:02     url2    2       url1
cookie1 2019-04-10 10:03:04     1url3   3       url1
cookie1 2019-04-10 10:10:00     url4    4       url1
cookie1 2019-04-10 10:50:01     url5    5       url1
cookie1 2019-04-10 10:50:05     url6    6       url1
cookie1 2019-04-10 11:00:00     url7    7       url1
cookie2 2019-04-10 10:00:00     url11   1       url11
cookie2 2019-04-10 10:00:02     url22   2       url11
cookie2 2019-04-10 10:03:04     1url33  3       url11
cookie2 2019-04-10 10:10:00     url44   4       url11
cookie2 2019-04-10 10:50:01     url55   5       url11
cookie2 2019-04-10 10:50:05     url66   6       url11
cookie2 2019-04-10 11:00:00     url77   7       url11

LAST_VALUE

取分组内排序后,截止到当前行,最后一个值

SELECT cookieid,
createtime,
url,
ROW_NUMBER() OVER(PARTITION BY cookieid ORDER BY createtime) AS rn,
LAST_VALUE(url) OVER(PARTITION BY cookieid ORDER BY createtime) AS last1 
FROM zcg1234;


cookieid  createtime            url    rn       last1  
-----------------------------------------------------------------
cookie1 2019-04-10 10:00:00     url1    1       url1
cookie1 2019-04-10 10:00:02     url2    2       url2
cookie1 2019-04-10 10:03:04     1url3   3       1url3
cookie1 2019-04-10 10:10:00     url4    4       url4
cookie1 2019-04-10 10:50:01     url5    5       url5
cookie1 2019-04-10 10:50:05     url6    6       url6
cookie1 2019-04-10 11:00:00     url7    7       url7
cookie2 2019-04-10 10:00:00     url11   1       url11
cookie2 2019-04-10 10:00:02     url22   2       url22
cookie2 2019-04-10 10:03:04     1url33  3       1url33
cookie2 2019-04-10 10:10:00     url44   4       url44
cookie2 2019-04-10 10:50:01     url55   5       url55
cookie2 2019-04-10 10:50:05     url66   6       url66
cookie2 2019-04-10 11:00:00     url77   7       url77

如果不指定ORDER BY,则默认按照记录在文件中的偏移量进行排序,会出现错误的结果

SELECT cookieid,
createtime,
url,
FIRST_VALUE(url) OVER(PARTITION BY cookieid) AS first2  
FROM zcg1234;

cookieid  createtime            url     first2
----------------------------------------------
cookie1 2019-04-10 10:00:02     url2    url2
cookie1 2019-04-10 10:00:00     url1    url2
cookie1 2019-04-10 10:03:04     1url3   url2
cookie1 2019-04-10 10:50:05     url6    url2
cookie1 2019-04-10 11:00:00     url7    url2
cookie1 2019-04-10 10:10:00     url4    url2
cookie1 2019-04-10 10:50:01     url5    url2
cookie2 2019-04-10 10:00:02     url22   url22
cookie2 2019-04-10 10:00:00     url11   url22
cookie2 2019-04-10 10:03:04     1url33  url22
cookie2 2019-04-10 10:50:05     url66   url22
cookie2 2019-04-10 11:00:00     url77   url22
cookie2 2019-04-10 10:10:00     url44   url22
cookie2 2019-04-10 10:50:01     url55   url22

SELECT cookieid,
createtime,
url,
LAST_VALUE(url) OVER(PARTITION BY cookieid) AS last2  
FROM zcg1234;

cookieid  createtime            url     last2
----------------------------------------------
cookie1 2019-04-10 10:00:02     url2    url5
cookie1 2019-04-10 10:00:00     url1    url5
cookie1 2019-04-10 10:03:04     1url3   url5
cookie1 2019-04-10 10:50:05     url6    url5
cookie1 2019-04-10 11:00:00     url7    url5
cookie1 2019-04-10 10:10:00     url4    url5
cookie1 2019-04-10 10:50:01     url5    url5
cookie2 2019-04-10 10:00:02     url22   url55
cookie2 2019-04-10 10:00:00     url11   url55
cookie2 2019-04-10 10:03:04     1url33  url55
cookie2 2019-04-10 10:50:05     url66   url55
cookie2 2019-04-10 11:00:00     url77   url55
cookie2 2019-04-10 10:10:00     url44   url55
cookie2 2019-04-10 10:50:01     url55   url55

如果想要取分组内排序后最后一个值,则需要变通一下:

SELECT cookieid,
createtime,
url,
ROW_NUMBER() OVER(PARTITION BY cookieid ORDER BY createtime) AS rn,
LAST_VALUE(url) OVER(PARTITION BY cookieid ORDER BY createtime) AS last1,
FIRST_VALUE(url) OVER(PARTITION BY cookieid ORDER BY createtime DESC) AS last2 
FROM zcg1234 
ORDER BY cookieid,createtime;

cookieid  createtime            url     rn     last1    last2
-------------------------------------------------------------
cookie1 2019-04-10 10:00:00     url1    1       url1    url7
cookie1 2019-04-10 10:00:02     url2    2       url2    url7
cookie1 2019-04-10 10:03:04     1url3   3       1url3   url7
cookie1 2019-04-10 10:10:00     url4    4       url4    url7
cookie1 2019-04-10 10:50:01     url5    5       url5    url7
cookie1 2019-04-10 10:50:05     url6    6       url6    url7
cookie1 2019-04-10 11:00:00     url7    7       url7    url7
cookie2 2019-04-10 10:00:00     url11   1       url11   url77
cookie2 2019-04-10 10:00:02     url22   2       url22   url77
cookie2 2019-04-10 10:03:04     1url33  3       1url33  url77
cookie2 2019-04-10 10:10:00     url44   4       url44   url77
cookie2 2019-04-10 10:50:01     url55   5       url55   url77
cookie2 2019-04-10 10:50:05     url66   6       url66   url77
cookie2 2019-04-10 11:00:00     url77   7       url77   url77

提示:在使用分析函数的过程中,要特别注意ORDER BY子句,用的不恰当,统计出的结果就不是你所期望。其实我觉得像这些场景都是比较特殊的,可以收藏一下,然后等啥时候遇到的业务场景或者情况需要复杂的逻辑的时候,普通的无法处理的时候,你就可以过来查询一下开窗函数的特性,有没有类似的功能。有的话,花个五分钟研究下,用就好了。如果不能用的话,那你就用UDF函数搞吧。如果再不行,写代码吧老哥....

你可能感兴趣的:(大数据,Hive)