spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Apache Spark (JIRA)" <j...@apache.org>
Subject [jira] [Assigned] (SPARK-16220) Revert ShowFunctions/ListFunctions in 2.0 to Reflect 1.6 Functionality
Date Mon, 27 Jun 2016 04:08:52 GMT

     [ https://issues.apache.org/jira/browse/SPARK-16220?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Apache Spark reassigned SPARK-16220:
------------------------------------

    Assignee:     (was: Apache Spark)

> Revert ShowFunctions/ListFunctions in 2.0 to Reflect 1.6 Functionality
> ----------------------------------------------------------------------
>
>                 Key: SPARK-16220
>                 URL: https://issues.apache.org/jira/browse/SPARK-16220
>             Project: Spark
>          Issue Type: Bug
>          Components: Spark Core, SQL
>    Affects Versions: 2.0.0, 2.0.1, 2.1.0
>            Reporter: Bill Chambers
>
> After discussing this with [~marmbrus] and [~rxin]. We've decided to revert SPARK-15663.
After doing some research it seems like this is an unnecessary departure from 1.X functionality
and does not have a reasonable substitute that gives the same functionality.
> The first step is to revert the change. After doing that there are a couple of different
ways to approachs to getting at user defined functions.
> 1. SHOW FUNCTIONS (shows all of them) + SHOW USER FUNCTIONS (Snowflake does this)
> 2. SHOW FUNCTIONS + SHOW USER FUNCTIONS + SHOW ALL FUNCTIONS
> 3. SHOW FUNCTIONS + SHOW SYSTEM FUNCTIONS (or something similar)
> 4. SHOW FUNCTIONS + some column to designate if it's system designed or user defined.
> 1. This aligns with previous functionality and then supplements it with something a bit
more specific. 
> 2. Is unclear because "all" is just unclear why does the default refer to only user defined
functions. This doesn't seem like the right approach.
> 3. Same kind of issue, I'm not sure why the user functions should be the default over
the system functions. That doesn't seem like the correct approach.
> 4. This one seems nice because it kind of achieves #1, keeps existing functionality,
but then supplants it with some more. This also allows you, for example, to create your own
set of date functions and then search them all in one go as opposed to searching system and
then user functions. This would have to return two columns though, which could potentially
be an issue?



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org


Mime
View raw message