[
https://issues.apache.org/jira/browse/DRILL-4482?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15185364#comment-15185364
]
Jason Altekruse commented on DRILL-4482:
----------------------------------------
[~acmeguy] I'm trying to reproduce this issue and not seeing it on a small avro
file. There is no guarantee about read order when reading a directory, so
running a limit 0 query over the same table in two formats (or even the same
list of files two different times) will not be guaranteed to give the same
result. Is transactions a directory or file without an extension?
Are you sure that there are not null values in this column? Could you try to
run a query with a predictable result like a max/min on the column or a limit
with a sort?
It is still possible that this is a Drill bug, and I will try with a
distributed query to see if I can reproduce it, but if you have time to try to
confirm any of these things it could help with creating a reproduction.
> Avro no longer selects data correctly from a sub-structure
> ----------------------------------------------------------
>
> Key: DRILL-4482
> URL: https://issues.apache.org/jira/browse/DRILL-4482
> Project: Apache Drill
> Issue Type: Bug
> Components: Storage - Avro
> Affects Versions: 1.6.0
> Reporter: Stefán Baxter
> Assignee: Stefán Baxter
> Priority: Blocker
> Fix For: 1.6.0
>
>
> Parquet:
> 0: jdbc:drill:zk=local> select s.client_ip.ip from
> dfs.asa.`/processed/<>/transactions` as s limit 1;
> +----------------+
> | EXPR$0 |
> +----------------+
> | 87.55.171.210 |
> +----------------+
> 1 row selected (1.184 seconds)
> Avro:
> 0: jdbc:drill:zk=local> select s.client_ip.ip from
> dfs.asa.`/streaming/<>/transactions` as s limit 1;
> +---------+
> | EXPR$0 |
> +---------+
> | null |
> +---------+
> 1 row selected (0.29 seconds)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
