| From: | "Bo?ena Potempa" <Bozena(dot)Potempa(at)otc(dot)pl> |
|---|---|
| To: | "Hiroshi Inoue" <Inoue(at)tpf(dot)co(dot)jp> |
| Cc: | <pgsql-odbc(at)postgresql(dot)org> |
| Subject: | Re: PostgreSQL+ Beta bug? |
| Date: | 2002-11-20 16:48:42 |
| Message-ID: | MKEIIFDEFPOPHIPFIFJCKEFKDHAA.Bozena.Potempa@otc.pl |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-odbc |
[..]
>According to an ODBC related document, the behavior of ODBC
>functions in such case is undefined and it is driver-specific
>if a driver handles this correctly.
OK. If it is driver-specific then why don't allow users to
to use this feature if it is implemented in the server?
>> insert/read data entirely making usage of provided length.
>> Is it implemented like that?
>
>No and how can I implement it ?
>Must I convert '\0' to "|| chr(0) ||" where there's no
>guarantee that it works ?
Why not?
Your statement suggests that parameter values are hard-coded
into SQL statements by ODBC driver. Is that true?
If you encode 0 as "|| chr(0) ||" it will change nothing:
if the server is able to store something after 0 then
it will be stored, otherwise not. If you afraid then
maybe this behaviour can be implemented as an option
to the ODBC data source (like "LF<->CR/LF Conversion")?
Is it much work to add such option?
Best regards,
Bozena Potempa
P.S.
0s in strings are important for some of our customers,
storing very strange things in their old Clipper
applications.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Robert John Shepherd | 2002-11-20 20:01:00 | Re: FATAL 1: Sorry, too many clients already |
| Previous Message | Cherubini Enrico | 2002-11-20 14:10:24 | data type conversion ? |