[ 
https://issues.apache.org/jira/browse/HIVE-6232?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Eugene Koifman updated HIVE-6232:
---------------------------------

    Description: 
Pig values support wider range than Hive.  e.g. Pig BIGDECIMAL vs Hive DECIMAL. 
 When storing Pig data into Hive table, if the value is out of range there are 
2 options:
1. throw an exception.
2. write NULL instead of the value

The 1st has the drawback that it may kill the process that loads 100M rows 
after 90M rows have been loaded.  But the 2nd may not be appropriate for all 
use cases.

Should add support for additional parameters in HCatStorer where the user can 
specify an option to controll this.

see org.apache.pig.backend.hadoop.hbase.HBaseStorage for examples

  was:
Pig values support wider range than Hive.  e.g. Pig BIGDECIMAL vs Hive DECIMAL. 
 When storing Pig data into Hive table, if the value is out of range there are 
2 options:
1. throw an exception.
2. write NULL instead of the value

The 1st has the drawback that it may kill the process that loads 100M rows 
after 90M rows have been loaded.  But the 2nd may not be appropriate for all 
use cases.

Should add support for additional parameters in HCatStorer where the user can 
specify an option to controll this.


> allow user to control out-of-range values in HCatStorer
> -------------------------------------------------------
>
>                 Key: HIVE-6232
>                 URL: https://issues.apache.org/jira/browse/HIVE-6232
>             Project: Hive
>          Issue Type: Sub-task
>          Components: HCatalog
>    Affects Versions: 0.13.0
>            Reporter: Eugene Koifman
>            Assignee: Eugene Koifman
>
> Pig values support wider range than Hive.  e.g. Pig BIGDECIMAL vs Hive 
> DECIMAL.  When storing Pig data into Hive table, if the value is out of range 
> there are 2 options:
> 1. throw an exception.
> 2. write NULL instead of the value
> The 1st has the drawback that it may kill the process that loads 100M rows 
> after 90M rows have been loaded.  But the 2nd may not be appropriate for all 
> use cases.
> Should add support for additional parameters in HCatStorer where the user can 
> specify an option to controll this.
> see org.apache.pig.backend.hadoop.hbase.HBaseStorage for examples



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

Reply via email to