Abnormal behaviour of Advanced Search on customized metadata

[edwin law]

Hi everyone

- We have added a customized metadata block "CityUHK" with a metadata field "RGCFunded". The metadata field is a controlled vocabulary with allowed values "Yes" or "No".

- There are four datasets with RGCFunded = "Yes" and two datasets = "No"

- We tried Advanced Search for RGCFunded = "Yes" and the result is correct.

- However, when we tried Advanced Search for RGCFunded = "No", no results were shown (should be 2 datasets).

- The tsv file for the customized metadata block is as follows:

Thanks

Edwin

[Philip Durbin]

Hi Edwin,

Thanks for letting us know about this. I poked around on Harvard Dataverse and I think I'm able to replicate the bug.

This dataset at https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/29136 , for example, has a mix of "Yes" and "No" answers to questions under its "Political Science Replication Initiative Metadata" metadata block, which is defined at https://github.com/IQSS/dataverse/blob/v6.6/scripts/api/data/metadatablocks/customPSRI.tsv

A search for "PSRI1:Yes" finds the dataset (and other datasets): https://dataverse.harvard.edu/dataverse/harvard/?q=PSRI1%3AYes

A search for "PSRI10:No" doesn't find that dataset (or any dataset) but should: https://dataverse.harvard.edu/dataverse/harvard?q=PSRI10:No

I'm not purposefully putting an URL-encoded colon in the "Yes" and a non-URL-encoded colon in the "No". I'm just pasting the URLs that Dataverse produces after doing a search. I also tried variations where I encoded the colon, and added quotes.

I'm not sure what's going on here. Please feel free to create an issue at https://github.com/IQSS/dataverse/issues

Thanks,

Phil

--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dataverse-community/011ad9c2-fb26-470c-b049-ec1a39aec807n%40googlegroups.com.

--

Philip Durbin
Software Developer for http://dataverse.org
http://www.iq.harvard.edu/people/philip-durbin

[edwin law]

Hi Philip

It's great that you can replicate the bug. BTW, our Dataverse version is 6.0, and the build no. is 1512-366fd41. Many thanks.

Regards

Edwin

[Philip Durbin]

Sure and thanks for creating https://github.com/IQSS/dataverse/issues/11621 about this.

To view this discussion visit https://groups.google.com/d/msgid/dataverse-community/c923c21f-bce0-4b64-8540-2311931cc294n%40googlegroups.com.