Problem with import leading to failed deduplication

[rootsandberries]

This is the same problem as issue #91 (which I mis-described as an import problem, so creating a new more aptly named issue here). @LukasWallrich you said you fixed it, but when I updated the packed on my computer and ran the assessing databases vignette, I'm now getting zero overlap for PsychInfo, where before I was seeing overlap locally, but not in the vignette online.

[LukasWallrich]

At least we are now getting consistency ;/

Could you identify a pair of records that are wrongly not recognised as overlapping?

[rootsandberries]

I'm attaching a csv file with the PsycInfo records that overlap with Web of Science (both records contained in the file--Zotero is identifying these as duplicates). I should also mention that Trevor is running into something similar where a ton of records from Dimensions that should be overlapping are not being identified as duplicates...he may post a separate issue!

psyc_wos_duplicates.csv

[TNRiley]

this was indeed related to #99 it appears that the raw dimensions .ris and the raw psycinfo .ris both had some type of issue. By importing the .ris into endnote first, then simply exporting it, the issue resolves. We still need to understand why this happens with specific metadata - potentially write a fix in for it OR a check/warning.

[LukasWallrich]

Reopened and renamed this - will have a look now.

[LukasWallrich]

Please ignore my comment - just saw that Trevor replaced the problematic file. Will check with the previous one.

[LukasWallrich]

It looks like I reintroduced a bug previously fixed in synthesisr - really not sure where I even got the old code version from. But this is fixed now, and works with psycinfo.

@TNRiley can you please check whether it also works with the Dimensions file you had trouble with? Then we can actually close this :)

[TNRiley]

@LukasWallrich I'm now getting hung up on read_citations, it starts to run but just sits there, I'll check back later to see if there is an error, currently just running forever

[TNRiley]

Looks good! - new impart ran a bit longer, but the fix looks good!