[erlang-questions] ODBC param query unicode

[Ryan Huffman]

Hi Ken,

Thank you for the suggestion.  We threw around a few ideas, including yours,
but ended up opting to put up an http interface which didn't take long.

I agree, the odbc driver is useful, but doesn't seem to get as much
attention as the other db-specific drivers (mysql, postgres).  I have to
imagine the relative demand in the erlang community is unfortunately low in
this regard.

Thanks,

Ryan

On Wed, Feb 23, 2011 at 3:04 PM, Ken Ellis <kaellis@REDACTED> wrote:

> Ryan,
>
> If you're thinking of throwing in the towel, have you looked at
> jinterface + JDBC?  We wound up ditching the ODBC library and using
> erlang:open_port to spawn a shell command.  But we're mostly doing
> bulk inserts, so COPY form stdin is fast and easy anyway.  We're also
> using a shell command to execute and parse the response from an SQL
> query.  I had considered briefly jinterface + JDBC to solve some
> issues, but did not come to any firm conclusions. Same boat as you,
> relatively exotic commercial db, nothing native.  The ODBC library is
> useful, for sure, but the level of finish is a bit lower than the rest
> of the libs in the OTP distribution.
>
> Ken
>
>
> On Wed, Feb 23, 2011 at 2:27 PM, Ryan Huffman <ryanhuffman@REDACTED>
> wrote:
> >>
> >> should use mysql native driver, It works fine.
> >
> >
> > Unfortunately I'm using SQL Server and not MySQL, thank you for the
> > suggestion though.
> >
> > It says add some unicode support!  (Some is an importante word here).
> >
> >
> > Hi Ingela,
> >
> > Sorry, I should have worded my question better.  I was mostly interested
> in
> > finding out what the extent of the added support was :).
> >
> > Thanks for the tip on the test cases.  Unfortunately I still haven't had
> any
> > luck.  The SELECTs still aren't returning unicode, just latin1 binaries.
> I'm
> > using Arch Linux, unixODBC and freeTDS.  Using wvarchar always fails with
> a
> > bind error - looking at the c code, it looks like a unixodbc or freetds
> > error.  At this point I'm about ready to throw the towel in.
> >
> > Thanks for the help,
> >
> > Ryan
> >
> > On Wed, Feb 23, 2011 at 1:42 AM, Ingela Andin <ingela.andin@REDACTED
> >wrote:
> >
> >> Hi!
> >>
> >> It says add some unicode support!  (Some is an importante word here).
> >> You have to
> >> handle the encoding yourself but you can insert and extract unicode
> >> data if you know
> >> how your datbase engine handles unicode data.
> >> See the test case unicode in odbc_data_type_SUITE.erl
> >>
> >> Regards Ingela Erlang/OTP team - Ericsson AB
> >>
> >> 2011/2/23 Ryan Huffman <ryanhuffman@REDACTED>:
> >> > According to the release notes the 2.10.8 added unicode support, but
> thus
> >> > far I haven't had any luck getting it to work (
> >> >
> >>
> http://www3.erlang.org/documentation/doc-5.8/lib/odbc-2.10.8/doc/html/notes.html
> >> ).
> >> >  Can anyone shed some light on what is supported?  I've tried using
> >> > param_query's with sql_wvarchar but without success (using utf8, utf16
> >> > little endian, utf16 big endian, etc...), and selecting rows with utf
> >> > characters returns a ? in place of utf characters (as expected, the
> rest
> >> of
> >> > the characters are still 8 bits wide rather than 16).
> >> >
> >> > Using 2.10.8 and R14B, connecting to an MSSQL server.
> >> >
> >> > Thanks,
> >> >
> >> > Ryan
> >> >
> >>
> >
>