Subject: Re: range_agg
To: Corey Huinker <corey.huinker@gmail.com>
Cc: Pgsql Hackers <pgsql-hackers@postgresql.org>
References: <16d71dc8-34cf-5ebd-1ce5-ccd93c0a14f9@illuminatedcomputing.com>
 <CADkLM=dP8f8T0oDxOsYWDacZO9ydj3o5xgr7aTvhZgQWonm3Bw@mail.gmail.com>
From: Paul Jungwirth <pj@illuminatedcomputing.com>
Message-ID: <91d7296b-7174-c2bb-8efa-7bfab53a97c4@illuminatedcomputing.com>
Date: Mon, 6 May 2019 11:17:46 -0700
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.6.1
MIME-Version: 1.0
In-Reply-To: 
 <CADkLM=dP8f8T0oDxOsYWDacZO9ydj3o5xgr7aTvhZgQWonm3Bw@mail.gmail.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 5/4/19 3:11 PM, Corey Huinker wrote:
>     One question is how to aggregate ranges that would leave gaps and/or
>     overlaps. So in my extension there is a one-param version that forbids
>     gaps & overlaps, but I let you permit them by passing extra parameters,
>     so the signature is:
> 
> 
> Perhaps a third way would be to allow and preserve the gaps.

Thanks for the feedback! I think this is what I'm doing already 
(returning an array of ranges), but let me know if I'm misunderstanding. 
My extension has these signatures:

     range_agg(anyrange) returning anyrange
     range_agg(anyrange, bool) returning anyarray
     range_agg(anyrange, bool, bool) returning anyarray.

The first variant raises an error if there are gaps or overlaps and 
always returns a single range, but the other two return an array of ranges.

I was planning to use the same signatures for my patch to pg, unless 
someone thinks they should be different. But I'm starting to wonder if 
they shouldn't *all* return arrays. I have two concrete use-cases for 
these functions and they both require the array-returning versions. Is 
it helpful to have a version that always returns a single range? Or 
should I make them all consistent?

Thanks,

-- 
Paul              ~{:-)
pj@illuminatedcomputing.com