[
https://issues.apache.org/jira/browse/CARBONDATA-2501?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16482392#comment-16482392 ]
Kevin Kong commented on CARBONDATA-2501:
----------------------------------------
Total data volume is 20 million, and the base number of "recitients" is 10 million
> When using presto to query the dictionary index column, presto server OOM
> -------------------------------------------------------------------------
>
> Key: CARBONDATA-2501
> URL:
https://issues.apache.org/jira/browse/CARBONDATA-2501> Project: CarbonData
> Issue Type: Bug
> Environment: presto 0.187
> carbondata in master branch/1.3.1/1.4
> Reporter: Kevin Kong
> Priority: Blocker
> Labels: Presto, dictionary
>
> # create tables sql:
> carbon.sql("CREATE TABLE IF NOT EXISTS pt2.order_detail ( id BIGINT, order_code STRING, sales_area_id INT, sales_id INT, order_inputer INT, pro_type STRING, currency INT, exchange_rate DECIMAL, unit_cost_price DECIMAL, unit_selling_price DECIMAL, order_num INTEGER, order_amount DECIMAL, order_discount DOUBLE, order_account_amount DECIMAL, order_time TIMESTAMP, delivery_channel INT, delivery_address STRING, recipients STRING, contact STRING, delivery_date DATE, comments STRING ) STORED BY 'carbondata' TBLPROPERTIES ('DICTIONARY_INCLUDE' = 'recipients,contact,sales_area_id,sales_id', 'NO_INVERTED_INDEX' = 'id,order_code' )").show()
>
> 2. presto query:
> when 'select id from carbondata.pt2.order_detail limit 1' is normal
> when 'select recipients from carbondata.pt2.order_detail limit 1' , presto server is oom.
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)