On Tue, Mar 30, 2010 at 19:32, Michael Bayer <mike...@zzzcomputing.com> wrote: > Bo Shi wrote: >>> pep 249 specifies "list of tuples" for fetchmany() and fetchall() >> >> Hrm, pep-249 seems to only specify "sequence" and "sequence of >> sequences" for the fetch*() functions, specifying list of tuples only >> as one possible example. Perhaps the C implementation of RowProxy is >> being too strict here? I'm surprised that pyodbc is the only dbapi >> implementation that this problem has occurred in... do all the other >> implementations subclass tuple for their rows? > > we run the tests all the time with Pyodbc, so I wasn't aware this was a > pyodbc issue. I'd run without the c extensions for now. For our C > extension to coerce into a tuple begins to add overhead and defeat the > purpose of using the extensions in the first place, though Gaetan would > have to answer this question.
Supporting arbitrary sequences is cheaper than I expected (you pay more than previously only if the sequence is neither a tuple nor a list), so now we are still fast for usual DBAPIs and hopefully don't break on odd stuff (as long as it is a sequence). -- Gaëtan de Menten -- You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalch...@googlegroups.com. To unsubscribe from this group, send email to sqlalchemy+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en.