Unclear how to find more granularity of files beyond "File Type" (application, tabulardata, data, etc.)

[pdurbin]

If you got to https://dataverse.harvard.edu/dataverse/polbehavior and click "Files" you see "File Type" as in the screenshot below, but the detail is not very granular. You see files flagged as application, tabulardata, data, etc.:

What if you want a specific type of file such as an R file? It's possible but you have to know which undiscoverable field to search on, which in this case is fileContentType. Specifically, a search for fileContentType:type/x-r-syntax ( https://dataverse.harvard.edu/dataverse/harvard?q=fileContentType%3Atype%2Fx-r-syntax ) will show 1,553 as of this writing in the Harvard Dataverse:

fileContentType (and related fields, possibly) should be exposed on the Advanced Search page, which currently looks like this:

The way I figured out to look for type/x-r-syntax is by looking at a list of files for a dataset in JSON format with this: https://dataverse.harvard.edu/api/datasets/:persistentId?persistentId=doi:10.7910/DVN/SCN9LA

The background on this issue is that this morning @christophergandrud were discussing how one would identify all the R files in an installation of Dataverse. As a starting point, we looked at http://dx.doi.org/10.7910/DVN/SCN9LA (screenshot below) which I happened to know has an R file called "Thal PB Replication.R" that can be used to reproduce plots in a paper. (Once this platform for reproducibility has been publicly announced, I'll mention it. 😄 )

Anyway, with fileContentType:type/x-r-syntax as a starting point, one should be able to iterate through all these R files via Search API: http://guides.dataverse.org/en/4.6/api/search.html#iteration

(This issue is related: What are the allowed search fields for the Search API q parameter? #2558).

[christophergandrud]

Thanks so much for putting this detailed not together @pdurbin!

[pdurbin]

It was a good exercise for me to write up some thoughts here but this issue is referenced from #2822 and #4377 so we can find it again, if necessary (or through search, of course). Until users ask for this, I think this issue is just noise among 850 other open issues so I'm closing it. 😄