Hi -
I have this doubt.
Why do i need to use an INSERT INTO .....
can I just create hdfs directories and map it to a hive external table setting
the location of the hdfs directories.
will this work ? please advise.
Thanks,
Raj
On Monday, November 4, 2013 8:34 AM, Matouk IFTISSEN
<matouk.iftis...@ysance.com> wrote:
Yes it is possible:
hadoop fs -mkdir /hdfs_path/'cust_id>1000'
I tested it and works, then you can store data in this directory .
for concat function you do simple:
insert into your_table_partionned PARTITION (path_xxx)
select attr,id, concat ('/data1/customer/', id) as path_xxx from your_table
where id <1000
......
Cdt.
2013/11/4 Raj Hadoop <hadoop...@yahoo.com>
How can i use concat function? I did not get it. Can you please elaborate.
>
>
>My requirement is to create a HDFS directory like
>(cust_id>1000 and cust_id<2000)
>
>
>
>and map this to a Hive External table.
>
>
>can i do that?
>
>
>
>On Monday, November 4, 2013 3:34 AM, Matouk IFTISSEN
><matouk.iftis...@ysance.com> wrote:
>
>Hello
>You can use concat function or case to do this like:
>Concat ('/data1/customer/', id)
>.....
>Where id <1000
>Etc..
>Hope this help you ;)
>Le 3 nov. 2013 23:51, "Raj Hadoop" <hadoop...@yahoo.com> a écrit :
>
>All,
>>
>>
>>I want to create partitions like the below and create a hive external table.
>>How can i do that ?
>>
>>
>>/data1/customer/id<1000
>>/data1/customer/id>1000 and id < 2000
>>
>>/data1/customer/id >2000
>>
>>
>>
>>Is this possible ( < and > symbols in folders ?)
>>
>>
>>My requirement is to partition the hive table based on some customer id's.
>>
>>
>>Thanks,
>>Raj
>
>
--
Matouk IFTISSEN | Consultant BI & Big Data
24 rue du sentier - 75002 Paris - www.ysance.com
Mob : +33 6 78 51 18 69 || Fax : +33 1 73 72 97 26
Ysance sur :Twitter | Facebook | Google+ | LinkedIn | Newsletter
Nos autres sites : ys4you | labdecisionnel | decrypt
