From: | Ron <ronljohnsonjr(at)gmail(dot)com> |
---|---|
To: | pgsql-general(at)lists(dot)postgresql(dot)org |
Subject: | Re: COMMIT IN STORED PROCEDURE WHILE IN A LOOP |
Date: | 2022-10-19 14:44:26 |
Message-ID: | b4ed6642-8717-35b0-5067-3bebab131f66@gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On 10/19/22 08:06, Mladen Gogala wrote:
[snip]
>
>>
>> Applications do this kind of thing all the time, very successfully; it
>> was just that the loop was in the application rather than in the procedure.
>>
>> High commit rates happen all the time, and they don't break PostgreSQL.
>> For example, an IoT application collecting sensor data and doing many
>> inserts per second is also doing many commits per second, since each bare
>> INSERT is in its own transaction. PostgreSQL handles it just fine.
>
> Point of my post is that the business logic, in your case it's IoT
> sensors, determines what is transaction and when to commit. Advice like
> "commit often and commit early", to paraphrase the famous Chicago mayor,
> is easy to find but I would take it with grain of salt.
In the normal course of operation (i.e, not when bulk loading), you /should/
commit at the end of every "business transaction". We've committed after X
business when running stovepipe "batch" jobs processing input files. In
those cases, though, we had to track progress through the file; in the case
of a rollback, the application had to go back to the last input file "save
point" and start over.
--
Angular momentum makes the world go 'round.
From | Date | Subject | |
---|---|---|---|
Next Message | Ron | 2022-10-19 14:49:26 | Re: Speeding up adding fky on a very large table |
Previous Message | Daniel Verite | 2022-10-19 14:29:24 | Re: How to store "blobs" efficiently for small and large sizes, with random access |