hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Adam Kramer (JIRA)" <>
Subject [jira] Commented: (HIVE-315) remove map() and reduce()
Date Fri, 29 May 2009 00:40:45 GMT


Adam Kramer commented on HIVE-315:

Another option is to actually fix them to ensure mapper/reducer semantics.

It would be lovely, for example, if MAP forced the data transformation to occur in a map step
(creating a new mapreduce process if necessary) and REDUCE forced the data transformation
to occur in a reduce step.

This would be very helpful for specifying how data is to be transformed, as MAP scripts are
inherently different from REDUCE scripts, since one may need to assume sorted data and the
other by definition cannot.

But if we can't make MAP and REDUCE actually refer/ensure that the program will be run in
the MAP or REDUCE step, I agree that this syntax should be eliminated because it is misleading.

> remove map() and reduce()
> -------------------------
>                 Key: HIVE-315
>                 URL:
>             Project: Hadoop Hive
>          Issue Type: Bug
>            Reporter: Venky Iyer
> this is a minor thing, but since these aren't actually doing anything to ensure mapper/reducer
semantics, they are misleading. everything is a transform. 

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message