Re: No duplicates on multiple colums when creating table

Richard, AH, now we have the correct question to answer..... Look in the syntax at 'primary key'.... a part of the create table statement, after the column definitions: create table xxx( a ..., b.... , c...., primary key (a,b,c) ).... THAT should fix your problem. Doug "Richard Krenek" <rkrenek@ihs.com> wrote in message news:39411BF7.DFD3206A@ihs.com... > The answer to the Ingres creating and index is yes as does Informix (behind > the scenes as an internal index). I know it is possible to do what I want in > the CREATE TABLE statement, it says so in the manual, but it doesn't tell > you how, it assumes I know what I'm doing. It assumed wrong. > > Richard Krenek > > Doug Agnew wrote: > > > Richard, > > > > I think you would have to use an index (and I suspect OpenIngres probably > > did so under the covers). Can you imagine searching a million row table > > SEQUENTIALLY every time you wanted to change a row (or insert a new one)? > > > > Doug > > "Richard Krenek" <rkrenek@ihs.com> wrote in message > > news:39410B9F.707267B9@ihs.com... > > > Hello all, > > > When doing a CREATE TABLE sql statment how do you declare a record > > > has to be unique? I know how to say a column has to be distinct but > > > cannot figure how to set a set of columns to be distinct. I can do it > > > via an index but I'd prefer to do it while creating the table. Here is > > > the (OpenIngres) SQL statment I'm trying to reproduce: > > > > > > create table doc_bib( > > > doc_global_id char(16) not null default ' ', > > > uis_item_time char(16) not null default ' ', > > > fam_uis_item_time char(16) not null default ' ', > > > ) > > > with noduplicates;> > > > > > That "with no duplicates" is throwing me a curve ball!!! > > > > > > Richard Krenek > > > >