Dataloader is not associating all records to a tearsheet correctly
BullhornUKTier3 opened this issue · 1 comments
Description
Associating records with a tearsheet does not add all records
Steps to Reproduce
- Run the first spreadsheet to create the tearsheet/ alternatively login and create the tearsheet manually
- Run the second spreadsheet to associate the Candidates with the newly created tearsheet
- Observe all records correctly added according to Dataloader
- Login to the system and observe that only some records have actually been associated
Expected behavior:
The expectation is for all records to be correctly associated with the tearsheet
Actual behavior:
Only a fraction of the records from the csv are correctly associated. I believe this to be being caused by the duplicate checker as viewing the calls within ELK you can see that Dataloader is firing off DELETE calls after the PUT calls which immediately unassociates the record with the tearsheet.
This does not occurr when the duplicate checker is disabled
Reproduces how often:
Everytime, through tests of associating 7 candidates/Contacts with a tearsheet the most common result was one record being associated
Additional Information
- The
dataloader.properties
file (minus login info) - The CSV input file (smallest possible file that reproduces the issue)
- The results file(s)
- The log file
--- All of this information is in the attached ZIP (Dataloader Files)
Dataloader files.zip
I just want to second this. Had the same issue the other day.
Here's the workaround I found (not ideal, but it does the job):
- Find out the actual ID of the tearsheet you need to add people to
- Do candidate export with the IDs of those you want to add to the tearsheet with the field 'tearsheet.id' as the second column for the export
- Open the results file and add your new tearsheet id to the existing ones in excel / CSV editor with a ; separator.
- Reimport the candidates using that newly created CSV.
This does accomplish the goal and does allow you to add candidates to a tearsheet. It's not ideal, but it works fine.