[
https://issues.apache.org/jira/browse/HIVE-1994?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12994581#comment-12994581
]
Adam Kramer commented on HIVE-1994:
-----------------------------------
Agree; also consider deprecating DISTRIBUTE/SORT/CLUSTER BY in favor of
DISTRIBUTED/SORTED/CLUSTERED BY, a syntax that would explicitly prevent
short-circuiting and subdivision for only the query it's a part of.
I can't imagine that "sort by in the subquery leads to assumptions in the
parent query" scales well or will last long in any case, but this functionality
is not only necessary for backwards-compatibility, but is also kind of the
entire reason Hive was developed and/or conceived: To utilize mapreduce
functionality in order to transform and process data. Preventing the querier
from making mapreduce assumptions just seems sad.
> Support new annotation @UDFType(stateful = true)
> ------------------------------------------------
>
> Key: HIVE-1994
> URL: https://issues.apache.org/jira/browse/HIVE-1994
> Project: Hive
> Issue Type: Improvement
> Components: Query Processor, UDF
> Reporter: John Sichi
> Assignee: John Sichi
>
> Because Hive does not yet support window functions from SQL/OLAP, people have
> started hacking around it by writing stateful UDF's for things like
> cumulative sum. An example is row_sequence in contrib.
> To clearly mark these, I think we should add a new annotation (with separate
> semantics from the existing deterministic annotation). I'm proposing the
> name stateful for lack of a better idea, but I'm open to suggestions.
> The semantics are as follows:
> * A stateful UDF can only be used in the SELECT list, not in other clauses
> such as WHERE/ON/ORDER/GROUP
> * When a stateful UDF is present in a query, there's an implication that its
> SELECT needs to be treated as similar to TRANSFORM, i.e. when there's
> DISTRIBUTE/CLUSTER/SORT clause, then run inside the corresponding reducer to
> make sure that the results are as expected.
> For the first one, an example of why we need this is AND/OR short-circuiting;
> we don't want these optimizations to cause the invocation to be skipped in a
> confusing way, so we should just ban it outright (which is what SQL/OLAP does
> for window functions).
> For the second one, I'm not entirely certain about the details since some of
> it is lost in the mists in Hive prehistory, but at least if we have the
> annotation, we'll be able to preserve backwards compatibility as we start
> adding new cost-based optimizations which might otherwise break it. A
> specific example would be inserting a materialization step (e.g. for global
> query optimization) in between the DISTRIBUTE/CLUSTER/SORT and the outer
> SELECT containing the stateful UDF invocation; this could be a problem if the
> mappers in the second job subdivides the buckets generated by the first job.
> So we wouldn't do anything immediately, but the presence of the annotation
> will help us going forward.
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
