Re: Can I do this in Query

Very clever, Evan!  Use an unmatched join!

Only thing is I think he'll have to define two fields for this new join
file, and it will give him more flexibility in sort capabilities, as
I'll explain.

The first field I think will have to define the 3-character match field
for two reasons:  1) I don't think Query will allow a substring spec on
the match, and 2) what if there's 'DAB', 'DWY', etc. that he doesn't
want to pool with 'DSL'?

The second field could be used to sequence the level breaks.  Say he
wants a sequence of the 'X' group first, the 'everything else' group
second, and the 'DSL' group last.  Assuming that a non-match would
generate a blank (x'40') value return to the query, the 'X' record
could have a sequence value of something less than x'40', and the 'DSL'
record could have a sequence value of something higher than x'40'.
Then the query can sort & level break on this sequence value.

I don't have a need for this at the moment, and am not absolutely
certain that this would work, so if someone tries it, please post back.

- Dan

--- Evan Harris <spanner@ihug.co.nz> wrote:
> Mark
>
> this (or a variation) might be a plan but I need to check the join
> properties to see if it is actually possible and can't get at a
> system
> right now....
>
> - create a file that has a one character field in it.
> - add a "D" and an "X" record
> - join the files (if possible) by substringing the first position
> - include unmatched items
> - break on the value in the joined to file (i.e. the single character
> field)
>
> Hope this helps
>
> Regards
> Evan Harris

__________________________________________________
Do you Yahoo!?
Yahoo! Web Hosting - Let the expert host your site
http://webhosting.yahoo.com