You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@carbondata.apache.org by "anubhav tarar (JIRA)" <ji...@apache.org> on 2018/04/06 15:43:00 UTC
[jira] [Comment Edited] (CARBONDATA-2282) presto carbon does not
support reading specific partition on which query is fired
mapreduce.input.carboninputformat.partitions.to.prune property is null
[ https://issues.apache.org/jira/browse/CARBONDATA-2282?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16428463#comment-16428463 ]
anubhav tarar edited comment on CARBONDATA-2282 at 4/6/18 3:42 PM:
-------------------------------------------------------------------
[~chenliang613]
In this Issue we are handling the scenario of querying data to standard hive partitioned table in carbondata.
although you are getting the correct data from partition table but in current master branch we were iterating over all the partitions to get the data not the one which is required,which will be slow you can check value of property mapreduce.input.carboninputformat.partitions.to.prune in carboninputformat class after executing query on a particular partition it is null so when getPartitionsToPrune method is called it is returning null value instead of required partition specs
but now in this pr we have identified the required partitions, so the query will be performed only on the required partition.
Firstly we are identifying the partition columns through domain constraints in presto and after that we are adding the Partition Spec to the configuration object to query partition
was (Author: anubhavtarar):
[~chenliang613]
In this PR we are handling the scenario of querying data to standard hive partitioned table in carbondata.
although you are getting the correct data from partition table but in current master branch we were iterating over all the partitions to get the data not the one which is required,which will be slow you can check value of property mapreduce.input.carboninputformat.partitions.to.prune in carboninputformat class after executing query on a particular partition it is null so when getPartitionsToPrune method is called it is returning null value instead of required partition specs
but now in this pr we have identified the required partitions, so the query will be performed only on the required partition.
Firstly we are identifying the partition columns through domain constraints in presto and after that we are adding the Partition Spec to the configuration object to query partition
> presto carbon does not support reading specific partition on which query is fired mapreduce.input.carboninputformat.partitions.to.prune property is null
> --------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: CARBONDATA-2282
> URL: https://issues.apache.org/jira/browse/CARBONDATA-2282
> Project: CarbonData
> Issue Type: Bug
> Components: core, presto-integration
> Affects Versions: 1.3.0
> Reporter: zhangwei
> Assignee: anubhav tarar
> Priority: Major
> Fix For: 1.3.0
>
> Attachments: partitonToPrune.patch
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)