There is a behavior in the Delimited Text driver that I need to turn off
if it's possible, it's very strange, but then again so am I.

I have a driver called HR Import Match who's job it is to import a
file. That part is working peachy. It has a series of match rules to
match up the user, with the expectation that some will be misses. It's
working, in my test data, 4 of 5 match, one record misses.

In the command transform, there is a rule which sends the event back to
the source command processor, so that the same record gets written back
out to an output file for a human to process and fix the directory. That
works AOK as well.

When you take that same file and move it back to the input, it
processes, and if it's still a miss, it sends it back to the source
command processor as another do-over. This is where things go awry
slighly. The trace shows it being processed, but no file is generated to
the output directory.

If, before re-submitting the file by moving it over, I add a blank
space to the end of the line (thereby changing it), it does write out
another output file.

So, it seems like there is *something* that is supressing the output of
an identical record on the subscriber but it seems far fetched that the
engineers would cache a prior record and supress it if it's a dupe, but
if that's what's hapenning, is there a way to stop it?

I did set allow-loopback=yes but that did not seem to have an effect on

rrawson's Profile:
View this thread: