From: | pabloa98 <pabloa98(at)gmail(dot)com> |
---|---|
To: | Joe Conway <mail(at)joeconway(dot)com> |
Cc: | "pgsql-generallists(dot)postgresql(dot)org" <pgsql-general(at)lists(dot)postgresql(dot)org> |
Subject: | Re: how to add more than 1600 columns in a table? |
Date: | 2019-04-24 22:11:35 |
Message-ID: | CAEjudX7BeWcL0xbYeTwVnBafVK3uHn0ABEkf7cgf8wE0MY6C9A@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Thank you Joe! I will take a look
Pablo
On Wed, Apr 24, 2019 at 1:47 PM Joe Conway <mail(at)joeconway(dot)com> wrote:
> On 4/24/19 4:17 PM, pabloa98 wrote:
> > Sadly today we hit the 1600 columns limit of Postgresql 11.
> >
> > How could we add more columns?
> >
> > Note: Tables are OK. We truly have 2400 columns now. Each column
> > represents a value in a matrix.
>
> As everyone else has mentioned, your use case sounds like arrays are
> appropriate. Note that PostgreSQL supports 2 dimensional (actually more
> than 2d if you wanted) arrays which are essentially perfect for
> representing a matrix.
>
> If this works for your data model it will likely save a ton of storage
> space and perform much better than alternatives.
>
> FWIW, if you are needing to do matrix math, you might want to look into
> PL/R (https://github.com/postgres-plr/plr) as it supports 2d arrays as
> arguments which are converted directly into R matrices.
>
> I don't know for sure but likely PL/Python could be used to process
> matrices as well.
>
> HTH,
>
> Joe
>
> --
> Crunchy Data - http://crunchydata.com
> PostgreSQL Support for Secure Enterprises
> Consulting, Training, & Open Source Development
>
>
From | Date | Subject | |
---|---|---|---|
Next Message | pabloa98 | 2019-04-24 22:14:13 | Re: how to add more than 1600 columns in a table? |
Previous Message | pabloa98 | 2019-04-24 22:11:04 | Re: how to add more than 1600 columns in a table? |