[
https://issues.apache.org/jira/browse/HIVE-1528?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Carl Steinbach updated HIVE-1528:
---------------------------------
Description:
Currently the only way to evaluate a path expression on a JSON object is
through get_json_object. If there are many fields in the JSON object need to be
extract, we have to call this UDF multiple times.
There are many use cases that get_json_object needs to be called many times in
one query to convert the JSON object to a relational schema. It would be much
desirable if we have a JSON UDTF that supports the following syntax:
{code}
select a.id, b.*
from a lateral view json_tuple(a.json_str, '$.f1', '$.f2', ..., '$.fn') b as
f1, f2, ..., fn
{code}
where the json_table function only scans the json_object once and return a set
of tuple of (f1, f2,..., fn).
was:
Currently the only way to evaluate a path expression on a JSON object is
through get_json_object. If there are many fields in the JSON object need to be
extract, we have to call this UDF multiple times.
There are many use cases that get_json_object needs to be called many times in
one query to convert the JSON object to a relational schema. It would be much
desirable if we have a JSON UDTF that supports the following syntax:
{code}
select a.id, b.*
from a lateral view json_table(a.json_object, '$.f1', '$.f2', ..., '$.fn') b
as f1, f2, ..., fn
{code}
where the json_table function only scans the json_object once and return a set
of tuple of (f1, f2,..., fn).
> Add json_tuple() UDTF function
> ------------------------------
>
> Key: HIVE-1528
> URL: https://issues.apache.org/jira/browse/HIVE-1528
> Project: Hive
> Issue Type: New Feature
> Components: UDF
> Affects Versions: 0.7.0
> Reporter: Ning Zhang
> Assignee: Ning Zhang
> Fix For: 0.7.0
>
> Attachments: HIVE-1528.2.patch, HIVE-1528.patch
>
>
> Currently the only way to evaluate a path expression on a JSON object is
> through get_json_object. If there are many fields in the JSON object need to
> be extract, we have to call this UDF multiple times.
> There are many use cases that get_json_object needs to be called many times
> in one query to convert the JSON object to a relational schema. It would be
> much desirable if we have a JSON UDTF that supports the following syntax:
> {code}
> select a.id, b.*
> from a lateral view json_tuple(a.json_str, '$.f1', '$.f2', ..., '$.fn') b as
> f1, f2, ..., fn
> {code}
> where the json_table function only scans the json_object once and return a
> set of tuple of (f1, f2,..., fn).
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
