Re: Normalize INTERVAL ouput format in a db driver

I could manage to identify/describe interval fields by testing the different values
returned by PQfmod().

Could someone confirm that PQfmod() returns will not change in future versions?

I have seen in the docs that there is a deprecated compile-time option that defines
how time, timestamp and intervals are stored (using double or int64), I guess this
should not affect the value returned by PQfmod()... right?

Thanks!
Seb

Sebastien FLAESCH wrote:
> Further, little libpq question:
>
> When using INTERVAL types, can I rely on PQfmod(), PQfsize() to determine
> the exact definition of the INTERVAL precision?
>
> => what YEAR/MONTH/DAY/HOUR/MINUTE/SECOND fields where used to create
> the column.
>
> I get different values for the type modifier, but how to interpret this?
>
> Can someone point me to the documentation or source code where I can find
> more details about this?
>
> I found this:
>
> http://www.postgresql.org/docs/8.4/static/libpq-exec.html#LIBPQ-EXEC-SELECT-INFO
>
>
> But there are not much details in PQfmod() description...
>
> Thanks!
> Seb
>
> Sebastien FLAESCH wrote:
>> Hi all,
>>
>> Just testing 8.4rc2 INTERVALs...
>>
>> According to the doc, INTERVAL output format is controlled by SET
>> intervalstyle.
>>
>> I am writing an interface/driver and need a solution to fetch/convert
>> interval
>> values independently from the current format settings...
>>
>> I could force my driver to implicitly set the intervalstyle to
>> iso_8601, but I
>> would prefer to leave this in the hands of the programmer...
>>
>> Imagine you have to write and ODBC interface/driver with libpq that
>> must support
>> the SQLINTERVAL C structure, how would you deal with PostgreSQL
>> intervals?
>>
>> Is it possible to query the current intervalstyle?
>>
>> Thanks!
>> Seb
>>
>>
>>
>
>