[
https://issues.apache.org/jira/browse/DRILL-7096?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16895733#comment-16895733
]
ASF GitHub Bot commented on DRILL-7096:
---------------------------------------
paul-rogers commented on issue #1829: DRILL-7096: Develop vector for canonical
Map<K,V>
URL: https://github.com/apache/drill/pull/1829#issuecomment-516246460
@KazydubB, thanks for the explanations. Probably not productive for us to
continue to debate the design now that the code is done. Probably best to get
this in so folks can use it; leaving refinements for later.
You correctly point out that EVF is used by only two readers, neither of
which support a TRUEMAP. The remaining EVF reader, JSON, uses true maps (not a
TRUEMAP) that can map string keys to any kind of value, so it will continue to
need to use the (non-true) MAP type. So, we won't need EVF TRUEMAP support for
that reader either. Instead, we can let a volunteer can add EVF support later
when it might be needed.
The thought about a new base class simply recognized that a TRUEMAP does not
have the same creation semantics as a Repeated MAP, so having TRUEMAP inherit
from the leaf Repeated MAP type introduces very odd code hacks. (Like saying a
Dog is a subclass of Cat, but without the purr and much more social.) But,
again, I suppose we can fix that later if we have the need.
Still, I would suggest that, if you can spare a bit of time, that we add
tests so we're not using the Parquet reader as our primary testing mechanism.
The Parquet reader is quite complex and debugging this code in the context of
the Parquet reader will be difficult.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
> Develop vector for canonical Map<K,V>
> -------------------------------------
>
> Key: DRILL-7096
> URL: https://issues.apache.org/jira/browse/DRILL-7096
> Project: Apache Drill
> Issue Type: Sub-task
> Reporter: Igor Guzenko
> Assignee: Bohdan Kazydub
> Priority: Major
> Fix For: 1.17.0
>
>
> Canonical Map<K,V> datatype can be represented using combination of three
> value vectors:
> keysVector - vector for storing keys of each map
> valuesVector - vector for storing values of each map
> offsetsVector - vector for storing of start indexes of next each map
> So it's not very hard to create such Map vector, but there is a major issue
> with such map representation. It's hard to search maps values by key in such
> vector, need to investigate some advanced techniques to make such search
> efficient. Or find other more suitable options to represent map datatype in
> world of vectors.
> After question about maps, Apache Arrow developers responded that for Java
> they don't have real Map vector, for now they just have logical Map type
> definition where they define Map like: List< Struct<key:key_type,
> value:value_type> >. So implementation of value vector would be useful for
> Arrow too.
--
This message was sent by Atlassian JIRA
(v7.6.14#76016)
