sqoop-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sowjanya Kakarala <sowja...@agrible.com>
Subject Re: Importing data to bucketed table in Hive
Date Wed, 28 Mar 2018 21:29:37 GMT
Thank you Douglas. Will definetly try it out. 
Before trying you saw my number of records right. Will this process minimizes the time?

My metrics for Sqoop to hive partitioned table and hive partitioned table to bucketing table,
took almost 24hrs for each table. (So little concerned about suggested scenario.)  

> On Mar 28, 2018, at 4:20 PM, Douglas Spadotto <dougspadotto@gmail.com> wrote:
> 
> Hi Sowjany,
> 
> You can try and import data into delimited files on hdfs and create an external table
that can be used to insert data into the bucketed table.
> 
> That is usually a pattern I use to load data into Hive. Only use managed tables for actual
user access.
> 
> Regards,
> 
> Douglas 
> 
> On Wednesday, March 28, 2018, Sowjanya Kakarala <sowjanya@agrible.com <mailto:sowjanya@agrible.com>>
wrote:
> Hi Guys,
> 
> I am trying to import the data from my db to Hive using Sqoop.
> 
> My usage is to get the data to Hive and later if necessary update the records from Hive.
For ACID transactions I learnt that I definitely need bucketing tables. 
> 
> Now I created bucketed table and I am trying to import the data using Sqoop import where
it is throwing a error as follows:
> 
> ERROR tool.ImportTool: Encountered IOException running import job: org.apache.hive.hcatalog.common.HCatException
: 2016 : Error operation not supported : Store into a partition with bucket definition from
Pig/Mapreduce is not supported 
> 
> If it is not supported is there a work around?
> 
> I created a partitioned table and import data into that table, can I insert these data
from partitioned table to created bucketing table? If yes can someone provide a best example?
> 
> I have tried with `insert into` statement for each partition which takes so long.
> Because my each year contains “241972735” records.
> 
> Also, this might be not a good practice for production environment.
> 
> Appreciate your help.
> 
> Thanks
> Sowjanya


Mime
View raw message