Data Edit API does not allow multiple ignored attributes

When editing ignore_attribute field with the data edit API with multiple attributes to ignore, only the last one is used:

import openml

new_id = openml.datasets.fork_dataset(41702)  # new ID was: 43070
openml.datasets.edit_dataset(new_id, ignore_attribute=["instance_id", "repetition", "runstatus"])

I verified that the Python API sends the following xml (indentation for readability):

<?xml version="1.0" encoding="utf-8"?>
<oml:data_edit_parameters xmlns:oml="http://openml.org/openml">
  <oml:ignore_attribute>instance_id</oml:ignore_attribute>
  <oml:ignore_attribute>repetition</oml:ignore_attribute>
  <oml:ignore_attribute>runstatus</oml:ignore_attribute>
</oml:data_edit_parameters>

however when obtaining the dataset XML after the edit:

<oml:data_set_description>
  <oml:id>43070</oml:id>
  <oml:name>MIP-2016-regression</oml:name>
  <oml:version>2</oml:version>
  <oml:description/>
  <oml:description_version/>
  <oml:format>ARFF</oml:format>
  <oml:upload_date>2019-05-28T13:09:37</oml:upload_date>
  <oml:licence>Public</oml:licence>
  <oml:url>
    https://www.openml.org/data/v1/download/21377444/MIP-2016-regression.arff
  </oml:url>
  <oml:file_id>21377444</oml:file_id>
  <oml:default_target_attribute>PAR10</oml:default_target_attribute>
  <oml:ignore_attribute>runstatus</oml:ignore_attribute>
  <oml:citation>[1] NA</oml:citation>
  <oml:visibility>public</oml:visibility>
  <oml:minio_url>
    http://openml1.win.tue.nl/dataset43070/dataset_43070.pq
  </oml:minio_url>
  <oml:status>active</oml:status>
  <oml:processing_date>2021-07-28 16:48:03</oml:processing_date>
  <oml:md5_checksum>91a74d1d325765e61530ebf8c2ba3263</oml:md5_checksum>
</oml:data_set_description>

changing the edit call again (e.g. to ["instance_id", "repetition"]) does change the dataset XML, so it's not a matter of waiting for updates. It consistently only saves the last field in the list.

Start with the openml.datasets.edit_dataset entry point and trace the data edit API request handling for repeated ignore_attribute elements. Reproduce the example with three attributes and compare the saved dataset XML with the request. Done means every requested ignored attribute is retained, rather than only the last one.