script hangs when mistakingly using Extractors.CSV instead of Extractors.TSV
Opened this issue · 3 comments
USQL script on ADLA were hanging indefinitely, until we realized that we had setup the wrong extractor (CSV instead of TSV) on a .tsv file.
Sure it is our fault, moreover because we specified silent: true.
Still, I would expect the script to complete (and not hang), but not extract any record (because they would all fail validation since there would be no comma on the rows).
What do you think?
Hmm... that is weird indeed. Do you have a joblink you could file an incident on so we could investigate why it was hanging? Normally a job should time out if a vertex does run longer than 5 hours.
Normally a job should time out if a vertex does run longer than 5 hours.
Oh ok maybe. But we have not let it run for so long...
Again: with the silent:true option I would expect the job to complete but not read any record:
I would expect the script to complete (and not hang), but not extract any record (because they would all fail validation since there would be no comma on the rows).
A job-id that demonstrates this behaviour: 9c0806e9-1e27-4286-8d0c-fd423430482c
Thanks... can you please send me the full job link in email to at microsoft dot com?