drill-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jacques Nadeau <jacq...@apache.org>
Subject Re: Querying relatively big GeoJSON data
Date Mon, 11 Aug 2014 03:36:36 GMT
Hey Luca,

This is great feedback.  The giant document wrapping thing is clearly a
common pattern and we're trying to figure out the right way to integrate.
Full JSON support is something we're committed to but we're not there
yet.   You've hit the two main things we're missing: single large json with
embedded records and heterogeneous arrays.  We'll have both addressed
shortly with the large single json probably coming first.   Can you file a
jira and I'll get one of our doc team guys to more clearly spell this out
in the documentation.

The url as data source is something we've discussed but haven't yet filed a
jira.   It would be pretty straight forward and is quite nice as you could
rely on mime type to understand format.   Would you be interested in trying
to put a prototype together to get it started? Even a good jira with an
outline of functionality and integration points would be very helpful.

Thanks again for the great feedback.

On Aug 10, 2014 7:20 PM, "Luca Morandini" <lmorandini@ieee.org> wrote:

> On 11/08/14 12:04, Ted Dunning wrote:
>> But I expect that the data that you have here can be profitably analyzed
>> by
>> Drill already by allowing this bit of un-nesting and not being quite so
>> doctrinaire about JSON purity.
> Hmm... it is sensible advice, but part of the appeal of Drill lies the "no
> ETL needed" claim.
> Since we are at that, let me abuse of your patience once more: can an URL
> act as data source (I am thinking about using CouchDB as source) ?
> Regards,
> Luca Morandini
> Data Architect - AURIN project
> Melbourne eResearch Group
> Department of Computing and Information Systems
> University of Melbourne
> Tel. +61 03 903 58 380
> Skype: lmorandini

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