Groups keyboard shortcuts have been updated
Dismiss
See shortcuts

Simple Presto / ORC question

72 views
Skip to first unread message

Ben Johnson

unread,
Feb 12, 2016, 4:20:56 PM2/12/16
to qubole...@googlegroups.com
Hi, I seem to be having trouble querying orc tables using presto. I have 2 tables in my account: raw_logs and raw_logs_orc. I was testing partitioning and loaded data into both tables. Unfortunately I'm not getting any results for the following query in *presto*. It is working fine in hive however: SELECT * from raw_logs_orc; The table definition is: CREATE EXTERNAL TABLE raw_logs_orc ( message Varchar(1000) ) PARTITIONED BY ( application_id INT, year INT, month INT, day INT ) STORED AS ORC LOCATION 's3n://tera-emr/raw_logs_orc/' TBLPROPERTIES ("orc.compress"="SNAPPY") ; Am I missing something basic here? Thanks for your help.

Ashish Dubey

unread,
Feb 12, 2016, 4:54:28 PM2/12/16
to Qubole Public Forum
Ben, 

We looked into your recent queries and it seems SELECT * succeeded.. Is this still an issue?

thanks,
Ashish

On Fri, Feb 12, 2016 at 1:20 PM, Ben Johnson <b...@terastream.io> wrote:
Hi, I seem to be having trouble querying orc tables using presto. I have 2 tables in my account: raw_logs and raw_logs_orc. I was testing partitioning and loaded data into both tables. Unfortunately I'm not getting any results for the following query in *presto*. It is working fine in hive however: SELECT * from raw_logs_orc; The table definition is: CREATE EXTERNAL TABLE raw_logs_orc ( message Varchar(1000) ) PARTITIONED BY ( application_id INT, year INT, month INT, day INT ) STORED AS ORC LOCATION 's3n://tera-emr/raw_logs_orc/' TBLPROPERTIES ("orc.compress"="SNAPPY") ; Am I missing something basic here? Thanks for your help.

--
You received this message because you are subscribed to the Google Groups "Qubole Public Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qubole-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/qubole-users/56be4cb4339eb_39003ffbf3f2ef346371e%40ip-10-145-33-241.mail.
For more options, visit https://groups.google.com/d/optout.

me...@qubole.com

unread,
Feb 16, 2016, 1:48:11 PM2/16/16
to Qubole Public Forum, b...@terastream.io
Hi Ben,

In presto default configuration, hive.metastore-cache-ttl is set to 20min. The cache was not refreshed when you first run the presto query, due to which you got stale results.
If your data is likely to change quickly, you could set the value of hive.metastore-cache-ttl accordingly. You can override presto configuration on cluster configuration page, by adding following lines in "override presto configuration" text box.
catalog/hive.properties:

hive.metastore-cache-ttl=0m


For details on how to do this, you can follow the following document:

Thanks,
Megha
Reply all
Reply to author
Forward
0 new messages