[ 
https://issues.apache.org/jira/browse/IMPALA-447?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dan Hecht resolved IMPALA-447.
------------------------------
    Resolution: Not A Problem

Appears to be a config issue. Please reopen if you disagree and include the 
reasoning.

> DROP TABLE should check write permissions for trash if enabled
> --------------------------------------------------------------
>
>                 Key: IMPALA-447
>                 URL: https://issues.apache.org/jira/browse/IMPALA-447
>             Project: IMPALA
>          Issue Type: Bug
>          Components: Catalog
>    Affects Versions: Impala 1.0.1
>         Environment: CDM 4.6
> Impala 1.0.1 parcels
> CDH 4.30 parcels
> Debian 6
>            Reporter: Nicolas PHUNG
>            Priority: Minor
>              Labels: ramp-up
>
> I keep having this issue: Data in Table stored as Parquetfile / TextFile 
> still persist after being dropped (and refreshed). Meaning after the drop, I 
> can create a new Table with the same name and do a select, my data will 
> "reappear". Is it from my configuration or something ?
> Steps to reproduce:
> {noformat} 
> echo 'Refresh Hive/Impala tables and show existing tables'
> impala-shell -i node2.lan -r -q "show tables"
> echo 'Refresh Hive/Impala tables and show existing tables'
> impala-shell -i node2.lan -q "DROP TABLE IF EXISTS mergeTable"
> echo 'Propagate delete on all nodes'
> impala-shell -i node1.lan -q "refresh"
> impala-shell -i node3.lan -q "refresh"
> impala-shell -i node2.lan -r -q "show tables"
> impala-shell -i node2.lan -q "CREATE TABLE mergeTable LIKE table1 STORED AS 
> PARQUETFILE"
> impala-shell -i node2.lan -q "insert into table mergeTable select * from 
> table1"
> impala-shell -i node2.lan -q "insert into table mergeTable select * from 
> table2"
> echo 'Refresh Hive/Impala all the Impala nodes'
> impala-shell -i node2.lan -r -q "show tables"
> impala-shell -i node1.lan -q "refresh"
> impala-shell -i node3.lan -q "refresh"
> {noformat} 
> Do this twice or more, you'll see that the mergeTable keeps growing in count 
> (just a select (*) is enough to see it on the merge table) even if we drop it 
> and the beginning of each run.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Reply via email to