You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@carbondata.apache.org by "Prasanna Ravichandran (Jira)" <ji...@apache.org> on 2020/10/14 07:03:00 UTC
[jira] [Commented] (CARBONDATA-3807) Filter queries and projection
queries with bloom columns are not hitting the bloom datamap.
[ https://issues.apache.org/jira/browse/CARBONDATA-3807?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17213670#comment-17213670 ]
Prasanna Ravichandran commented on CARBONDATA-3807:
---------------------------------------------------
!bloom_issue_verification_after_load.PNG!
> Filter queries and projection queries with bloom columns are not hitting the bloom datamap.
> -------------------------------------------------------------------------------------------
>
> Key: CARBONDATA-3807
> URL: https://issues.apache.org/jira/browse/CARBONDATA-3807
> Project: CarbonData
> Issue Type: Bug
> Environment: Ant cluster - opensource
> Reporter: Prasanna Ravichandran
> Priority: Major
> Fix For: 2.0.0
>
> Attachments: bloom-filtercolumn-plan.png, bloom-show index.png
>
>
> Filter queries and projection queries with bloom columns are not hitting the bloom datamap.
> Bloom datamap is unused as per plan, even though created.
> Test queries:
> drop table if exists uniqdata;
> CREATE TABLE uniqdata (cust_id int,cust_name String,active_emui_version string, dob timestamp, doj timestamp, bigint_column1 bigint,bigint_column2 bigint,decimal_column1 decimal(30,10), decimal_column2 decimal(36,36),double_column1 double, double_column2 double,integer_column1 int) stored as carbondata;
> load data inpath 'hdfs://hacluster/user/prasanna/2000_UniqData.csv' into table uniqdata options('fileheader'='cust_id,cust_name,active_emui_version,dob,doj,bigint_column1,bigint_column2,decimal_column1,decimal_column2,double_column1,double_column2,integer_column1','bad_records_action'='force');
> create datamap datamapuniq_b1 on table uniqdata(cust_name) as 'bloomfilter' PROPERTIES ('BLOOM_SIZE'='640000', 'BLOOM_FPP'='0.00001');
> show indexes on uniqdata;
> explain select count(*) from uniqdata where cust_name="CUST_NAME_00000"; --not hitting;
> explain select cust_name from uniqdata; --not hitting;
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)