missing records not recognised in asset import in Jira SM cloud

[Maurice Gisberts]

How does Jira SM cloud version work with import in Assets? 
I have made an import for assets in Jira SM. It works fine the first time, but when I delete the imported file and I run Import again wihtout changing anything the import wil read the file (that is no longer on my drive) anyway. Thas means that the imported file is stored somewhere else an dhe Import is not reading the file on my computer. 
When I change the import file in the import specifications to the same file that I replaced with another file with the same name as before, but with less records en some new records. The Jira Import will read the new file, but does nog recognise missing records. It will however recognise new records in the import and import them in the assets. So I have no possibility to do something with missing records. Does anybody knows the exact way how Jira SM cloud version works on this specific point?

[Tiago Paladino]

Hi @Maurice Gisberts ,

Two separate things going on here, let me untangle them.

On the "file still gets read after you delete it locally": that's actually expected. When you upload a CSV to Assets, it gets stored on Atlassian's side, not read from your computer each time. So deleting it locally does nothing. To bring in a new version, you have to open the import configuration and upload the new file there. Just replacing it on your drive won't do a thing.

On the missing records: that's a setting in the object type mapping, not a global thing. Open your import, expand the object type mapping, and look for a field called "Missing objects". You'll get three choices: Ignore, Flag as missing, or Remove. Default is usually Ignore, which is exactly what you're seeing. Change it to Flag as missing (safer, just marks them so you can review) or Remove (deletes them). If you go with Remove, also check the threshold, it's a safety net that stops the deletion if too many objects would disappear at once.

One extra gotcha to be aware of: the missing-object detection only works on objects that were originally created by that same import configuration. If some of your objects were made by hand or by a different import, JSM won't touch them even with Remove selected. Also, make sure your import has a proper unique identifier set (like serial number or asset tag), because that's how it decides "this row equals this existing object".

Try flipping Missing objects to Flag as missing first, run it, and see if the ones you expect show up as missing. If they do, you're good to switch to Remove.

[Maurice Gisberts]

Thanks for the quick answer. And you are right. It where two different questions. The first answer is clear. 
For the second one. I have imported a file with 6 records and the import settings:

Missing Objects -> Remove (we tried also the update version to update an attribute)
Treshold -> 20

Works great. Then I deleted 5 records form the import file and added 2 new records. I reselected the import file and started the import file again. It recognises the new records but does not recognises the 5 missing objects that are no longer in the file. I have tried it with the same file name, a changed file name. Results are the same. I understand that the comparison for recognising the missing records in the second import is with the last succesfully import (the file stored on Atlassian's side?), but it seems that if I reselect the import file that the comparison does not work. The import doesn't recognise the missing objects. 

That's why I try to understand how it should work in Jira SM cloud.