pg.ddx.io  pgsql-general@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Laurenz Albe <laurenz.albe@cybertec.at>
To: Lorusso Domenico <domenico.l76@gmail.com>
To: pgsql-general@lists.postgresql.org
Subject: Re: Array vs Temporary table vs Normal Table + truncate at end
Date: Tue, 05 Sep 2023 04:01:50 +0200
Message-ID: <aaa8a8491f8fc91694d3791c73f12ee846c62478.camel@cybertec.at> (raw)
In-Reply-To: <CAJMpnG7QakV4-XRggcjB289W0=C+jV8HNcrQXzzhPFDK7H=ADA@mail.gmail.com>
References: <CAJMpnG7QakV4-XRggcjB289W0=C+jV8HNcrQXzzhPFDK7H=ADA@mail.gmail.com>

On Mon, 2023-09-04 at 18:42 +0200, Lorusso Domenico wrote:
> I prepared a function that build a set of records of type (myComplexType).
> 
> That works, but I've to use this record set in different situations.
> 
> One of them require to scan multiple times the function results:
>    1. comparing with another table (myProducteAttributeTable) to determine if I need to delete old record
>    2. check if the recordset contains attribute never registered before
>    3. in this case insert in appropriate table (myAttribute), retrieve attribute uid and type
>    4. finally update first table (myProducteAttributeTable) with new information coming from the recordset (with a merge)
> 
> So, which is the best approach?
> 
>  * Store result in an array and use in each subsequent query the unnest (with a complex record type?)
>  * Create temporary table on commit delete, with the same structure? But in this case I can't declare a
>    cursor for the table, and I can't use update where current of (for point 3)
>  * Create normal table and truncate as part of execution, but what happens in case of parallel execution?

I'd say it depends.

If the set is rather smallish, arrays may be the best way.

If the set is large (or the rows themselves are large), and the function is not called all the time,
I'd probably go with a temporary table.

If the set is large and the function called all the time, I'd try to use a persistent table to
avoid catalog table bloat.  If several function invocations need to store their record set in
the same table, you could make the backend process ID part of the primary key.

Yours,
Laurenz Albe





view thread (2+ messages)

Message-ID: <aaa8a8491f8fc91694d3791c73f12ee846c62478.camel@cybertec.at>
Permalink:  ../aaa8a8491f8fc91694d3791c73f12ee846c62478.camel@cybertec.at/
Also on:    postgresql.org/message-id/aaa8a8491f8fc91694d3791c73f12ee846c62478.camel@cybertec.at

 · 

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients from your mail client; the link fills in
  the headers below: reply via email
  Some clients, webmail especially, drop In-Reply-To and References from
  a link, and the reply then starts a new thread: use the mbox below.

  To: pgsql-general@postgresql.org
  Cc: laurenz.albe@cybertec.at, domenico.l76@gmail.com
  Subject: Re: Array vs Temporary table vs Normal Table + truncate at end
  In-Reply-To: <aaa8a8491f8fc91694d3791c73f12ee846c62478.camel@cybertec.at>
  References: <CAJMpnG7QakV4-XRggcjB289W0=C+jV8HNcrQXzzhPFDK7H=ADA@mail.gmail.com> <aaa8a8491f8fc91694d3791c73f12ee846c62478.camel@cybertec.at>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox