[midPoint] CsvConnector 2.8 can't load users because of nullvalue
Gomringer Michael
Michael.Gomringer at ti8m.ch
Fri Jul 25 10:32:33 CEST 2025
Hi Markus
Many thanks for the hint. That was a test on my side, to see if the first of the fields needs to be taken, sorry for that. I need to go by "EMPL_ID" and I've reconfigured it accordingly, still the same error.
I found something interesting, on the resource objects in the resource, there's an entry "Fatal Error" which is not deletable, nor viewable, as it seems to be an object with no identifier. This object might lead to the mentioned error, because no other change in the mappings or csv-file changes anything on the behaviour.
Maybe you can help me to remove this object? A delete ends in this error:
"Could not delete null, reason: Cannot invoke "com.evolveum.midpoint.xml.ns._public.common.common_3.ObjectType.getEffectiveOperationPolicy()" because "object" is null"
I removed all resources and created a new csv-resource, the corrupt object is still there. Can't find it anywhere else, e.g. in the repository-object list. Do I have to remove it from the DB directly somehow?
Many thanks and regards,
gomi
Von: Markus Calmius <markus.calmius at proton.ch>
Gesendet: 25 July 2025 10:16
An: MidPoint General Discussion <midpoint at lists.evolveum.com>
Cc: Gomringer Michael <Michael.Gomringer at ti8m.ch>
Betreff: Re: [midPoint] CsvConnector 2.8 can't load users because of nullvalue
Hi Gomi,
I did see one thing:
in MidPoint the Name attribute has to be unique.
In the schema-handling you have this:
<attribute id="57">
<ref>ri:MandNr</ref>
<inbound id="107">
<name>Name</name>
<strength>normal</strength>
<target>
<path>name</path>
</target>
</inbound>
</attribute>
and in the csv they both have MandNr = 1000
Change the attribute-mapping or the value and see if that solves the issue.
Kind regards,
Markus
On Friday, 25 July 2025 at 08:54, Gomringer Michael via midPoint <midpoint at lists.evolveum.com> wrote:
Dear all
I'm struggling with a very simple CSV Connector 2.8 in midpoint 4.9.3.
Any import task ends in an error: "Value of attribute '__NAME__' must be a single value, but it has nullvalues".
The tolerance on the mappings is set to "undefined", as written in an older thread that it shouldn't be on "false" [1], and I tried it already with "true" but that didn't help.
The connection to the csv can be established and the schema can be loaded. I down stripped the CSV to two lines only, UTF-8 encoded, everything in quotes, etc.
The only difference to the training-setup is the version (2.8 now vs. 2.7. in training) where the configuration was very simple and worked directly.
Attached is the xml of the csv-connector as well as an anonymized version of the CSV which ends in the same error and a log from the midpoint-container of the import-task.
I'm new to midpoint and we're waiting for a subscription. Please tell me if I need to provide any additional information. Help much appreciated.
Many thanks and cheers,
Gomi
[1] https://lists.evolveum.com/pipermail/midpoint/2020-June/006291.html
[Das Bild wurde vom Absender entfernt. Ai Powered Innovation]<https://www.ti8m.com/>
[Ai Powered Innovation] <https://www.ti8m.com/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.evolveum.com/pipermail/midpoint/attachments/20250725/02a24dbd/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ~WRD0003.jpg
Type: image/jpeg
Size: 823 bytes
Desc: ~WRD0003.jpg
URL: <https://lists.evolveum.com/pipermail/midpoint/attachments/20250725/02a24dbd/attachment.jpg>
More information about the midPoint
mailing list