[
https://issues.apache.org/jira/browse/PHOENIX-2584?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
maghamravikiran updated PHOENIX-2584:
-------------------------------------
Attachment: PHOENIX-2484-2.patch
[~jamestaylor], [~prkommireddi]
I have made two changes to the earlier patch.
a) Determine the sql type of the column using ColumnProjector. This helps in
cases where a column is defined as VARCHAR type but the user uses a REGEX_SPLIT
function on that column in the SQL query passed to LOAD. This causes the
resultant data type to be of Array type.
b) Added two tests , one that tests Arrays in SQL query and the other where the
user specifies just the table name in LOAD statement.
Can you please have a review.
You will notice the order of imports in PhoenixRecordWritable have changed and
they match the order of the settings file phoenix.importorder that we have.
> Support Array datatype in phoenix-pig module
> --------------------------------------------
>
> Key: PHOENIX-2584
> URL: https://issues.apache.org/jira/browse/PHOENIX-2584
> Project: Phoenix
> Issue Type: Bug
> Reporter: maghamravikiran
> Assignee: maghamravikiran
> Fix For: 4.7.0
>
> Attachments: PHOENIX-2484-2.patch, PHOENIX-2584-1.patch,
> PHOENIX-2584.patch
>
>
> The plan is to map an array data type column to a Tuple in Pig.
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
