HIVE本身是不太支持更新的，要从其中删除某一行其实也是费劲的。
不过HIVE本身还提供一种机制，可以删除其中的分区。只要某一条记录在某个分区中，就可以实现用个“转弯”的方式来实现，即先删除分区，再手动去掉这条记录，再导入到分区中。
如首先查到某个表中有记录7904个，
用这个命令：ALTER TABLE shphonefeature DROP IF EXISTS PARTITION(year = 2015, month = 10, day = 1);删除掉指定分区
再一查数据，就没有任何数据了。
hive> select count(*) from shphonefeature; Query ID = ndscbigdata_20160331105618_575ad188-25b8-4de8-9f79-0aa306908193 Total jobs = 1 Launching Job 1 out of 1 Number of reduce tasks determined at compile time: 1 In order to change the average load for a reducer (in bytes): set hive.exec.reducers.bytes.per.reducer=<number> In order to limit the maximum number of reducers: set hive.exec.reducers.max=<number> In order to set a constant number of reducers: set mapreduce.job.reduces=<number> Starting Job = job_1459079550905_0023, Tracking URL = http://ubuntu-bigdata-5:8088/proxy/application_1459079550905_0023/ Kill Command = /usr/hadoop/bin/hadoop job -kill job_1459079550905_0023 Hadoop job information for Stage-1: number of mappers: 2; number of reducers: 1 2016-03-31 10:56:24,782 Stage-1 map = 0%, reduce = 0% 2016-03-31 10:56:30,096 Stage-1 map = 50%, reduce = 0%, Cumulative CPU 3.27 sec 2016-03-31 10:56:31,135 Stage-1 map = 100%, reduce = 0%, Cumulative CPU 5.53 sec 2016-03-31 10:56:34,234 Stage-1 map = 100%, reduce = 100%, Cumulative CPU 6.88 sec MapReduce Total cumulative CPU time: 6 seconds 880 msec Ended Job = job_1459079550905_0023 MapReduce Jobs Launched: Stage-Stage-1: Map: 2 Reduce: 1 Cumulative CPU: 6.88 sec HDFS Read: 516060 HDFS Write: 5 SUCCESS Total MapReduce CPU Time Spent: 6 seconds 880 msec OK 7910 Time taken: 17.572 seconds, Fetched: 1 row(s) hive> ALTER TABLE shphonefeature DROP IF EXISTS PARTITION(year = 2015, month = 10, day = 1); Dropped the partition year=2015/month=10/day=1 OK Time taken: 0.238 seconds hive> select * from shphonefeature limit 10; OK Time taken: 0.117 seconds
