Glen,  am I correct in infering that the frtarget test case (the first
one) works with the pipeline at Marist, but the second one fails?

If so can you please provide a test case with data?

   j.

On 14 October 2010 00:45, Glenn Knickerbocker <[email protected]> wrote:
> I don't know how I never stumbled on this before.  I was playing with
> CALLPIPE, COLLATE, LOOKUP, and GATE together in handling a log file with
> nonunique timestamps, and for the longest time I couldn't figure out why
> the record on the primary of GATE kept vanishing.  I kept trying
> different ways of hooking up COLLATE, which I've never used much, but it
> turned out LOOKUP was the culprit!  It was releasing the primary input
> record even if no corresponding record was released on the primary or
> secondary output.
>
> This seems to have been at least partly fixed at some point.  For
> example:
>
>  PIPE literal 1 2 3 3 4 5 | split | frtarget lookup autoadd | cons
>
> does write the second 3 using 1.1.12, but writes just 4 and 5 using
> 1.1.10.  But this simple CALLPIPE consumes an extra record even with the
> current version:
>
>  'callpipe (end /) *..0: | exact: lookup detail
>  | end: gate' ,
>                '/  *..1: | exact:               | end: | *..0:'
>
> ¬R
>

Reply via email to