Closed abigailalexander closed 2 years ago
Due to the fact that we cannot use instrument command programs to write data, we need to have some flexibility on the metadata generation to support the different workflows across our groups. This sometimes means having to append details to already generated datasets in Scicat. We would prefer in this scenario to use update_dataset
rather than to check , insert a new data set and then delete the previous.
I hope that the update_dataset
function would be a useful feature for other teams and won't break between the old and new versions.
Thanks for the PR! Since there has been no release of pyscicat since the upsert was added, I think it would be safe to delete rather than deprecate those methods.
Thanks for the quick response, I'll remove them!
Following on from
upsert_dataset
not being supported in the new backend of scicat, I want to instead propose anupdate_dataset
method that uses patch to update a dataset based on pid. This method would hopefully be compatible both with the current scicat and future scicat versions (taking a quick look at the new scicat backend, it looks like similar functionality still exists there using patch for updating datasets).