RE: Constraint comparing blanks.

["Dennis Lovelady" <iseries@xxxxxxxxxxxx>]

The moment I sent this I realized I'd been caught in the trap of confusion
wrought by too much DDS work lately.  Of course you meant DDL, as you said.
Sorry.  But I stand behind the rest of my message, and would very much like
to see more detail about that failure.  Maybe a 
    SELECT NAME FROM TABLE WHERE NAME = ' '
or some such.  Or to prove this to yourself:
    SELECT NAME, COUNT(*) 
    FROM TABLE 
    WHERE NAME = ' '
    GROUP BY NAME

> > When you are declare DDL for an SQL defined table we use a lot of
> stuff
> > like
> > this
> >
> > NAME <> ' ';
>
> You do this in DDL?  Can you give an example DDL statement where you do
> this?  (Maybe CREATE VIEW?)
>
> Anyway, you have it right.  NAME <> ' ' will always and forever expand
> the shorter field with blanks until the lengths match, and will then
> compare.  One blank is and has always been equal to n blanks in SQL
> unless something's broken.
>
>
> create table ab (nam  char(150))
> Table AB in DLOVELADY1 created but could not be journaled.
> insert into ab values(' THOMAS   ')
> 1 rows inserted in AB in DLOVELADY1.
> insert into ab values('                    ')
> 1 rows inserted in AB in DLOVELADY1.
> select count(*) from ab where nam <> ' '
> COUNT ( * )
>            1
> SELECT statement run complete.
> select count(*) from ab where nam = ' '
> COUNT ( * )
>            1
> SELECT statement run complete.
>
> Can you provide more detail on how this failed for you, please?

Dennis Lovelady
http://www.linkedin.com/in/dennislovelady
--
"Diplomacy is the art of saying 'Nice doggie!' till you can find a rock."
        -- Wynn Catlin