Bill:

Thank you. I'll run an experiment.

bruce

> -------- Original Message --------
> Subject: [RBASE-L] - Re: Inner Join: Joining 3 or more tables?
> From: Bill Downall <[email protected]>
> Date: Tue, November 16, 2010 11:43 am
> To: [email protected] (RBASE-L Mailing List)
> 
> 
> Bruce,
> 
> There's also the UNION ALL syntax, which basically stacks selects on top of
> each other. Be sure the first SELECT allows R:BASE to infer the exact
> correct data type for the ordinal position of each item selected (e.g. TEXT
> (20)), and use parentheses around columns in later selects that you want
> R:BASE to cast into the matching data type from above:
> 
> SELECT ClientID, InvoiceNumber, InvoiceTotal +
> FROM  Clients c1, Invoices i2 +
> WHERE c1.ClientID = i2.ClientID +
> UNION ALL +
> SELECT ClientID, NULL, (0) +
> FROM Clients c3 +
> WHERE NOT EXIST +
> SELECT clientID from Invoices i4 WHERE i4.ClientID = c3.ClientID) +
> UNION ALL +
> SELECT NULL, InvoiceNumber, InvoiceTotal +
> FROM Invoices i5 +
> WHERE i5.ClientID IS NULL
> 
> Bill
> 
> 
> 
> 
> On Tue, Nov 16, 2010 at 2:35 PM, Bruce Chitiea <[email protected]>wrote:
> 
> > Larry:
> >
> > Between this and the workaround, I'm in great shape.
> >
> > Thanks much.
> >
> > bruce
> >
> > > -------- Original Message --------
> > > Subject: [RBASE-L] - Re: Inner Join: Joining 3 or more tables?
> > > From: Lawrence Lustig <[email protected]>
> > > Date: Tue, November 16, 2010 11:21 am
> > > To: [email protected] (RBASE-L Mailing List)
> > >
> > >
> > > You can perform multiple INNER JOINs in a single statement using the
> > older
> > > SQL-89 syntax:
> > >
> > > SELECT * FROM table1, table2, table3 +
> > >   WHERE table1.colX = table2.colY AND table2.colA = table3.colB
> > >
> > > Unfortunately, this works only for implicit INNER JOINs in the older
> > syntax.
> > >  For SQL-92 syntax (which is the only way to do OUTER JOINs) you are
> > limited to
> > > 2 tables per SELECT statement.  The work-around there is to use nested
> > views as
> > > described in your message.
> > > --
> > > Larry
> >
> >
> >


Reply via email to