RE: SQL vs. traditional I/O?

> -----Original Message-----
> From: midrange-l-bounces@xxxxxxxxxxxx / rob@xxxxxxxxx
> Sent: Friday, July 23, 2004 3:04 PM
>
> That took me a minute.  Basically one field for every 'element'
> of every column.  And each row in this fref would be a column
> in another table.  Sounds a lot like QSYS2/SYSCOLUMNS.

No, my "data FRF" would still be used as the FRF is now, to reference fields
in the "FRF".  Fields in data files would themselves never be defined in the
FRF; ideally, all fields would reference the FRF.

> The biggest time I see FREF's getting full is when they do something like
> R FREFR
>   IPROD         15          TEXT('Item Number')
>                             COLHDG('Item Number')
>   SLPROD        15          TEXT('Item')
>                             COLHDG('Item')
> ...
> Not that this is any better
> R FREFR
>   IPROD         15          TEXT('Item Number')
>                             COLHDG('Item Number')
>   SLPROD           R        REFFLD(IPROD)
> ...
> Instead of just having IPROD in there and having all files call it
> IPROD.  And then your programs can either qualify, or prefix, the file if
> they just have to.  (Like SL is just so much more meaningful than I.)

Even worse, when IPROD & SLPROD are defined with 500 other fields between
them.

db