[jira] [Comment Edited] (PHOENIX-1) select only gives results for certain combinations of selected columns when performing join

["Sergio Esteves (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/PHOENIX-1?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13911735#comment-13911735 ] 

Sergio Esteves edited comment on PHOENIX-1 at 2/25/14 6:31 PM:
---------------------------------------------------------------

Actually I just started with a clean environment and instead of using the github mirror I pulled last version from apache repo, using the new package configuration. This issue got fixed somehow. 


was (Author: sesteves):
This issue is still occurring in my environment. I've been working now to enable fine-grained logging in phoenix on the server side.

> select only gives results for certain combinations of selected columns when performing join
> -------------------------------------------------------------------------------------------
>
>                 Key: PHOENIX-1
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-1
>             Project: Phoenix
>          Issue Type: Bug
>            Reporter: Sergio Esteves
>
> I'm using last version from master branch ( 995f508b2a80d9158467c708f35afcf6da4f0cce ) and I've been facing a strange behavior.
> When I run the following queries:
> bq. select * FROM customer inner join address ON customer.c_addr_id = address.addr_id  inner join country ON address.addr_co_id = country.co_id;
> bq. SELECT c_id, c_uname, c_passwd, c_fname, c_lname, c_addr_id, c_phone, c_email, c_since, c_last_login, c_login, c_expiration, c_discount, c_balance, c_ytd_pmt, c_birthdate, c_data, addr_id, addr_street1, addr_street2, addr_city, addr_state, addr_zip, addr_co_id, co_id, co_name, co_exchange, co_currency FROM customer inner join address ON customer.c_addr_id = address.addr_id  inner join country ON address.addr_co_id = country.co_id;
> the resulting table is empty. 
> But if I remove some columns, like this:
> bq. SELECT c_id, c_uname, c_passwd, c_fname, c_lname, c_addr_id, c_phone, c_email, c_since, c_last_login, c_login, c_expiration, c_discount, c_balance, c_ytd_pmt FROM customer inner join address ON customer.c_addr_id = address.addr_id  inner join country ON address.addr_co_id = country.co_id;
> bq. SELECT c_id, c_uname, c_passwd, c_fname, c_lname, c_addr_id, c_phone, c_email, c_since, c_last_login, c_login, c_expiration, c_discount, c_balance, c_ytd_pmt, addr_id FROM customer inner join address ON customer.c_addr_id = address.addr_id  inner join country ON address.addr_co_id = country.co_id;
> the resulting table is not empty anymore, listing all rows correctly. Seems to me that there is some sort of limit on the size (in bytes) that all aggregated values of a row in the result can have.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)