Re: FMTDTA

[Douglas Handy <dhandy1@xxxxxxxxxxxxx>]

Booth,
 
>Which is fastest when there's a need?  FMTDTA still blows away the new young
>bucks so far as I know.
> 
>yes, yes, the new stuff is good, no doubt about it, but lets be kind to the
>old methods that built the empire in the first place.

But Booth, ever since around V2Rx something or possibly V3R1 or so, you can have
the performance of FMTDTA and still have the ability to use field names (ala
OPNQRYF) rather than field offsets (ala FMTDTA).

All you need to so is specify OPNQRYF ... ALWCPYDTA(*OPTIMIZE) and the query
optimizer will dynamically determine whether or not the sort will be faster.  In
those cases, it will perform a sort exactly like FMTDTA.  But in cases where it
determines a pre-existing index will outperform the sort it will use it instead.
And yes, it also considers the record selection criteria.   So you get the best
of both worlds.

You get to use field names -- not record offsets -- while still getting the
performance of FMTDTA.  I haven't used FMTDTA in many years because of this
change.

Prompt the help text on OPNQRYF ALWCPYDTA(*OPTIMIZE).

Doug