1.支持limit n,m语法的hive版本
#!/bin/bash
cnt=$(set hive.cli.print.header=false;hive -e "select count(*) from exam.userbehavior;")
echo "总记录数: $cnt 条"
#每页10条
pageSize=10
pageNum=`expr $(($cnt + $pageSize - 1)) / $pageSize`
echo "总页数: $pageNum 页"
for((a=0;a<=$pageNum;a++))
do
x=`expr $a + 1`
echo "***********第$x页记录**********"
hive -e "select * from exam.userbehavior limit $a,$pageSize;"
done
2.不支持limit n,m使用row_number实现
这种方式使用row_number最好写一个临时表,不然每次都排序很耗性能
#!/bin/bash
cnt=$(set hive.cli.print.header=false;hive -e "select count(*) from exam.userbehavior;")
echo "总记录数: $cnt 条"
for((a=1;a<=$cnt;a+=10))
do
x=`expr $a + 9`
hive -e "
select *
from
(select *,row_number() over(order by user_id) as rn from exam.userbehavior) a
where a.rn between $a and $x
"
done