Db.define_table('foo',Field('a'),Field('b'), Field('ab',unique=True,
compute=lambda r: r.a + r.b))http://zerp.ly/rochacbruno Em 30/06/2011 23:04, "Nick Arnett" <[email protected]> escreveu: > On Thu, Jun 30, 2011 at 6:43 PM, Nick Arnett <[email protected]> wrote: > >> >> >> On Thu, Jun 30, 2011 at 4:15 PM, Anthony <[email protected]> wrote: >> >>> IS_NOT_IN_DB does take a DAL Set object as the first argument, so you can >>> limit the records checked to a particular set within the table (see >>> http://web2py.com/book/default/chapter/07#Database-Validators). However, >>> I'm not quite sure what you're trying to do here. >>> >>> I think what he wants is the equivalent of Django's "unique together", >> which I'm also trying to figure out. IS_NOT_IN_DB seems to accomplish it, >> but is that going to create a unique index on the combined fields? >> Otherwise, it will be expensive to do inserts. >> > > Actually, I don't see how NOT_IN_DB can accomplish this, since it is > associated with a single field. Do I have to create an extra concatenated > field and make it unique? Seems silly, given that's what a database unique > index is. > > In other words, for table foo, if I want fields A, B and C to be unique > together, do I have to do this? > > db.define_table('foo', Field('A'), Field('B'), Field('C'), > Field('ABC_unique', unique=True)) > > And remember to concat A, B, and C into that last field every time I do an > insert? > > I'm missing something here, since this is so simple elsewhere. > > Nick
