Thanks, my index work for partition is more clear now. I can just create a set of partitions for index table as the same as the original table. And add index data to corresponding partitions.
On 09-7-9 下午10:17, "Namit Jain" <nj...@facebook.com> wrote: > If a table is partitioned, we should only allow insertion/loading data in a > partition – not the table. > So, load data local inpath > \"/Users/char/Documents/workspace/Hive-Clean/data/files/kv1.txt\" into table > srcpart ; > should fail. > > I think there was already a jira for that – let me check and otherwise file > one. > > > Thanks, > -namit > > > On 7/9/09 6:13 AM, "He Yongqiang" <heyongqi...@software.ict.ac.cn> wrote: > >> Hi all, >> The situation: >> create table srcpart(key string, value string) partitioned by (ds string, hr >> int); >> load data local inpath >> \"/Users/char/Documents/workspace/Hive-Clean/data/files/kv1.txt\" overwrite >> into table srcpart PARTITION (ds='2008-04-08', hr=11); >> load data local inpath >> \"/Users/char/Documents/workspace/Hive-Clean/data/files/kv1.txt\" into table >> srcpart ; >> >> And the command ‘select * from srcpart’ only access the partition data? And >> the data outside any partition is left untouched? >> >> >> Yongqiang >
