Re: BUG #13920: pg_try_advisory_xact_lock bigint trouble

On Monday, February 8, 2016, David G. Johnston <david(dot)g(dot)johnston(at)gmail(dot)com>
wrote:

> On Friday, February 5, 2016, <mtakvel(at)gmail(dot)com
> <javascript:_e(%7B%7D,'cvml','mtakvel(at)gmail(dot)com');>> wrote:
>
>> The following bug has been logged on the website:
>>
>> Bug reference: 13920
>> Logged by: Valeriy
>> Email address: mtakvel(at)gmail(dot)com
>> PostgreSQL version: 9.5.0
>> Operating system: Ubuntu
>> Description:
>>
>> Hello, I have few high load big tables. My logic calls
>> pg_try_advisory_xact_lock(bitint) for locking row in current table. As I
>> see
>> with bigint param pg_try_advisory_xact_lock lock same ids for all my
>> tables.
>> Insthead lock only row in one current table. Looks like this is bug and
>> will
>> be cool if you fix it.
>>
>>
> Likely working as designed. If you wish to provide an example of what you
> are doing we can probably explain your misunderstanding. Basically,
> though, there is nothing about the ID you pass to the advisory lock
> functions that cause them to be associated with a table. The ID is simply
> a number. You should try the two-key version and associate the first key
> with the table (probably oid) and the second with the row on that table.
>
>
Though the two-arg uses Integer so maybe not...

You should poprobably explain your use case a bit on the -general list if
you'd like to discuss alternatives. But the behavior described is how
things work right now.

David J.