[APP大数据项目]第四天

已经注册了 getdaybegin,getweekbegin,getmonthbegin,formatime（根据时间ms转为日期的字符串）的UDF

Hive查询日新增、周新增、月新增

查询某app当日新增用户数

SELECT count(*) FROM (
SELECT min(createdatms) mintime 
FROM ext_startup_logs 
WHERE appid = 'sdk34734' 
GROUP BY deviceid  
HAVING mintime >= getdaybegin() AND mintime < getdaybegin(1)) t;

查询某app 指定日期的日新增用户数

SELECT count(*) FROM (
SELECT min(createdatms) mintime 
FROM ext_startup_logs 
WHERE appid = 'sdk34734' 
GROUP BY deviceid  
HAVING mintime >= getdaybegin('2020/03/27 00:00:00') AND mintime < getdaybegin('2020/03/28 00:00:00',1)) t;

查询某app 当前周新增

SELECT count(*) FROM (
SELECT min(createdatms) mintime 
FROM ext_startup_logs 
WHERE appid = 'sdk34734' 
GROUP BY deviceid  
HAVING mintime >= getweekbegin() AND mintime < getweekbegin(1)) t;

查询某app 上周新增

SELECT count(*) FROM (
SELECT min(createdatms) mintime 
FROM ext_startup_logs 
WHERE appid = 'sdk34734' 
GROUP BY deviceid  
HAVING mintime >= getweekbegin(-1) AND mintime < getweekbegin()) t;

查询某app 某天所在周新增

SELECT count(*) FROM (
SELECT min(createdatms) mintime 
FROM ext_startup_logs 
WHERE appid = 'sdk34734' 
GROUP BY deviceid  
HAVING mintime >= getweekbegin('2020/03/27 00:00:00') AND mintime < getweekbegin('2020/03/27 00:00:00',1)) t;

查询某app 当前月新增

SELECT count(*) FROM (
SELECT min(createdatms) mintime 
FROM ext_startup_logs 
WHERE appid = 'sdk34734' 
GROUP BY deviceid  
HAVING mintime >= getmonthbegin() AND mintime < getmonthbegin(1)) t;

查询某app总用户数

SELECT count(distinct(deviceid)) FROM ext_startup_logs

活跃数

查询日活跃用户数

运用分区表，减少不必要的检索

SELECT count(distinct deviceid ) FROM ext_startup_logs 
WHERE appid = 'sdk34734' AND 
ym = formattime(getdaybegin(),'yyyyMM') AND
day = formattime(getdaybegin(), 'dd');

月活周活相似

一次查询出一周内每天的日活跃数

SELECT formattime(createdatms,'yyyy/MM/dd') ,count(distinct(deviceid))
FROM ext_startup_logs
WHERE appid = 'sdk34734' AND createdatms >= getweekbegin() AND createdatms < getweekbegin(1)
GROUP BY formattime(createdatms,'yyyy/MM/dd') 

一次查询出一个月内每周的周活跃数

formattime UDF内重载了 (long, string ,int)的方法，其中最后一个参数并没有用上，至少为了重载该方法，用于实现获取某天周起始日期的字符串。
实际工程应该另写一个UDF，此处为了方便.

SELECT formattime(createdatms,'yyyy/MM/dd',0) ,count(distinct(deviceid))
FROM ext_startup_logs
WHERE appid = 'sdk34734' AND createdatms >= getmonthbegin() AND createdatms < getmonthbegin(1)
GROUP BY formattime(createdatms,'yyyy/MM/dd',0)