>>>>> "Evandro's" == Evandro's mailing lists (Please, don't send personal messages to this address) <listasjr@xxxxxxxxx> writes: [I would have replied to your personal address, but I'm not about to copy it from a footer.] Evandro's> I'm doing a PhD in data mining and I need more than 1600 columns. I got an Evandro's> error message saying that I can not use more than 1600 columns. Evandro's> It is happening because I have to change categorical values to binary Evandro's> creating new columns. Do you know if oracle can handle it? /me boggles You are doing a PhD in data mining, and you have a table that needs more than 1600 columns? /me gasps What are they *teaching* these days? If you have a design that has more than 20 or so columns, you're probably already not normalizing properly. There just aren't *that* many attributes of a object before you should start factoring parts of it out, even if it means creating some 1-1 tables. In programming, if I ever see someone name a sequence of variables, like "thing1" and "thing2", I know there's going to be trouble ahead, because that should have been a different data structure. Similarly, I bet some of your columns are "foo1" and "foo2". Signs of brokenness in the design. Or do you really have 1600 *different* attributes, none of which have a number in their name? That requires a serious amount of creativity. :) -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 <merlyn@xxxxxxxxxxxxxx> <URL:http://www.stonehenge.com/merlyn/> Perl/Unix/security consulting, Technical writing, Comedy, etc. etc. See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training! ---------------------------(end of broadcast)--------------------------- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
