hbase预分region方法

hbsae表使用VIN（车架号 每个车辆唯一的编号 17位）
作为rowkey，但是VIN分布并不均匀，因此在数据量很大的时候需要考虑自定义预分region间隔区间：

use asmp;
select sub,count(1) n from (select substring(vin,1,5) as sub from tt_repair_deed_tmp where partition_brand='vw') a
group by sub order by n desc

select sub,count(1) n from (select substring(vin,1,7) as sub from tt_repair_deed_tmp where partition_brand='vw') a
group by sub order by n desc

select sub,count(1) n from (select substring(vin,1,9) as sub from tt_repair_deed_tmp where partition_brand='vw') a
group by sub order by n desc

select sub,count(1) n from (select substring(vin,1,11) as sub from tt_repair_deed_tmp where partition_brand='vw') a
group by sub order by n desc

按照4位划分region


按照5位划分region

按照7位划分region

按照9位划分region

正常来说有多少region会产生多少reduce，如果reduce太多占用资源也会很多，因此选择按照5位划分region。

然后自定义vin_split文件