Re: DBD::ODBC and issues getting LongTruncOk / LongReadLen to take effect
Ron Savage <[email protected]> Fri, 16 Nov 2018 11:11:33 +1100
| Newsgroups | gmane.comp.lang.perl.modules.dbi.general |
|---|---|
| Message-ID | <[email protected]> |
Hi Dan Does it make any difference if you put LongReadLen and LongTruncOK next to RaiseError in the attrs? On 16/11/18 10:54, Daniel Kasak wrote: > Hi Martin. Sorry for the very long delay. We had abandoned the ODBC > driver in favour of the native DBD::DB2, and I've been working on this > project outside my normal work hours anyway, so got bogged down in other > issues ... > > I've uploaded the trace to: https://tesla.duckdns.org/downloads/trace.log > The script I'm using to generate this is: > https://tesla.duckdns.org/downloads/db2_syscat_columns.pl > > It dies on the 1st call of $sth->fetchrow_hashref(): > DBD::ODBC::st fetchrow_hashref failed: st_fetch/SQLFetch (long > truncated DBI attribute LongTruncOk not set and/or LongReadLen too > small) (SQL-HY000) at db2_syscat_columns.pl > <http://db2_syscat_columns.pl> line 36. > > Thanks for your time ... > > Dan > > On Thu, Apr 5, 2018 at 9:50 PM Martin J. Evans via dbi-users > <[email protected] <mailto:[email protected]>> wrote: > > On 05/04/18 12:24, Daniel Kasak wrote: > > Hi all. > > > > I'm writing a database utility that has to access IBM's "DashDB" > and other DB2-variants. I have their latest ODBC driver, and I have > simple queries working. However queries against their system catalog > are not working - queries appear to be returning *empty* recordsets. > After some laborious debugging, I can see that when I call > $sth->fetchrow_array, DBI::errstr is set to: > > > > st_fetch/SQLFetch (long truncated DBI attribute LongTruncOk not > set and/or LongReadLen too small) (SQL-HY000) > > > > I've dealt with this before. So when constructing my $dbh, I do: > > > > $self->{dbh}->{LongReadLen} = 65535 * 1024 # 64MB - yes I know > it's big > > $self->{dbh}->{LongTruncOK} = 1; > > > > This usually fixes things - or at least has for SQL Server and > Netezza. For DB2 connections however, it appears to have no effect. > > > > I've turned on ODBC tracing immediately prior to calling > $sth->prepare ... $sth->execute ... $sth->fetchrow_array: > > https://paste.pound-python.org/show/hS6ur7dwGRsQubr29HFT/ > > > > This one was from the query: > > select * from SYSCAT.COLUMNS > > > > Does anyone know why LongTruncOk / LongReadLen are not being > honored here? > > > > I'm using unixODBC-2.3.6, DBD::ODBC from git, and the latest > available DB2 ODBC driver. > > > > Please help! Thanks :) > > > > Dan > > There are some very long columns in that result-set but I can't tell > from the unixodbc log alone what has happened. > Can you reduce your perl to the simplest test you can and rerun with > > DBI_TRACE=15=trace.log perl mycode.pl <http://mycode.pl> > > and show me the trace.log file. > > Martin > -- Ron Savage - savage.net.au