I have a workflow (screenshot attached), whereby I'm getting an error on processing when run with AMP on, but no error when AMP is off. The error is Dynamic Input (29) Internal Error in RecordReader::ReadFields: Too many fields in the Dynamic Input tool. I'm feeding in a list of excel files/tabs/arrays to read in (attached screenshot). I read the data out for both instances, seeing that one row of data was dropped from two of the tabs of data when I see the AMP error (ends up being two rows each, since the last actual row is blanks across). These are not the tabs with the most data, so not sure why these are dropping and others aren't. The data on all of the tabs is consistent (e.g. same number of columns, same format --- all dates and numbers), so it's unclear to me why two tabs are dropping and others are not. Also, these are not the tabs with the most data/rows.
My current resolution seems to be turn off AMP, though requires every member of my team to do this, in which case I may ask everyone to always turn it off. Is there any other solution? Is there an explanation for this error occurring?
@changwen - turn off AMP. Leave it off. Always.
when turning it off, it would pop up a window when running the workflow that you to check the amp engine again.
click on do not show the message again. re-run. sleep better at night.
@changwen are you using the version 22.1 with Patch 10?
Are you able to try the workflow with version 23.2? Bug and crash reporting have been improved in 23.2 and it would provide much better insights.
still not working. same error - a bug
Are you writing to either a network drive or a database? sometimes the pipe ends when you begin with a connectoin (like a VPN) and then at some point your connection times out. For some DBs this can also happen if your cluster goes to sleep.
@changwen you did not mention the version you are using. Are you able to share workflow?
If you are using 23.2, please enable Crash report in user setting and run the workflow. You will get a report ID please share that.