Re: Re: any way for a transaction to "see" inserts done earlier in the transaction?

[Robert DiFalco <robert.difalco@xxxxxxxxx>]

Right. I don't know this code or DBI but many frameworks create a pool of ids using sequence generators so that they can minimize round trips and know the id of new records before the are written. 

Sent from my iPhone

On Apr 17, 2014, at 8:06 AM, Susan Cassidy <susan.cassidy@xxxxxxxxxxxxxxxxxxxxxxxx> wrote:

Except for the fact that I get the new id returned from the first insert, which means that the insert probably did happen. 

Susan

On Wed, Apr 16, 2014 at 11:55 PM, Alban Hertroys <haramrae@xxxxxxxxx> wrote:

On 17 Apr 2014, at 2:49, David G Johnston <david.g.johnston@xxxxxxxxx> wrote:

> Robert DiFalco wrote
>> Two common cases I can think of:
>>
>> 1. The PERL framework is only caching the insert and does not actually
>> perform it until commit is issued.
>
> Wouldn't the same mechanism cache the corresponding SELECT?

Not likely, or if it did it wouldn’t be able to know what id was returned from the function (which calls nextval(), but that isn’t relevant here since it’s marked volatile).
That makes it a possible scenario for what’s being witnessed here.

Alban Hertroys
--
If you can't see the forest for the trees,
cut the trees and you'll find there is no forest.



--
Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general