incubator-couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Samuel Williams <space.ship.travel...@gmail.com>
Subject Re: add_lib in reducers?
Date Thu, 04 Apr 2013 08:01:00 GMT
I see. There are many cases when reduce code may need library
functionality. One very important case is when you are dealing with custom
data types, e.g. decimal numbers, that require specific functions to add,
subtract, etc.

The entire query server interface seems like a huge mess. It would be nice
if there was some way to improve it e.g. CouchDB 2.0 could have an entirely
new query server interface.

I'd be happy to write a proposal but I wouldn't have time to dedicate to it
unless I knew it was going to be seriously considered.


On 4 April 2013 19:21, Alexander Shorin <kxepal@gmail.com> wrote:

> Hi Samuel!
>
> There is issue with patch about it:
> https://issues.apache.org/jira/browse/COUCHDB-1202
>
> And there is the explanation why so:
>
> > One other issue: currently I haven't yet implemented commonjs for
> > reduce functions. The reason I have not is that reduce functions are
> > not batched, so library code would have to be sent and eval'd every
> > single time the reduce function is run. On top of this, 99% of the
> > time the reduce function won't make use of the library code, so it
> > will just be dead weight. (I have a hard time thinking of when a
> > reduce function will need to require library code).
>
>
> http://mail-archives.apache.org/mod_mbox/couchdb-dev/201009.mbox/%3CAANLkTi=hFvB8__sSfqtfZykv3E-H0oDcduqaZMtVE9Rc@mail.gmail.com%3E
>
> Hope they make things clear (:
>
> --
> ,,,^..^,,,
>
>
> On Thu, Apr 4, 2013 at 5:44 AM, Samuel Williams
> <space.ship.traveller@gmail.com> wrote:
> > Hi,
> >
> > I've implemented a Ruby query server.
> >
> > I wanted to have libraries available to reduce functions.
> >
> > Is it okay to have add_lib add libraries to both mapper context and
> reduce
> > context?
> >
> > The method of loading libraries seems a bit cumbersome in the current
> > implementation, reducers are treated quite differently from mappers. Is
> > there any plan to improve this going forward?
> >
> > Kind regards,
> > Samuel
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message