RIS import, tags on seperte lines are not separated

sethmg · May 18, 2011

Hi,
I have been trying to import an RIS file exported from EndNote.

I have changed the incorrect AB tag to the correct N2 tag, and all seemed to go well. However, when I tried to sync online I got a tag-too-long error and found that the tags had not been imported correctly.

The original file contained

KW - Tag1
Tag2
Tag3
Tag4

In the end I got it to work by writing a script to make this

KW - Tag1
KW - Tag2
KW - Tag3
KW - Tag4

The RIS format specification in clear that EndNote's method is wrong, stating "Each keyword or phrase must be on its own line, preceded by this tag ['KW - ']." I know it sucks to bend the rules, but maybe zotero can be modified to be tolerant of this error?

dstillman · May 18, 2011

I haven't tested this, but in theory this should already work:

else if(tag == "KW") {
	// keywords/tags
	
	// technically, treating newlines as new tags breaks the RIS spec, but
	// it's required to work with EndNote
	item.tags = item.tags.concat(value.split("\n"));
}

Maybe it's due to Windows newlines? value.split(/[\n\r]{1,2}/)?

sethmg · May 18, 2011

Hi Dan
I thought about the new line markings too and used gedit to save a copy of the file with Windows new lines and a copy with Unix line endings. Both had the same behaviour.

I did not physically check that the line endings were properly Win/Unix, but I've never had gedit mess that up before.

dstillman · May 18, 2011

Sample record?

sethmg · May 18, 2011

Windows line endings: http://ubuntuone.com/p/uBA/
Unix line endings: http://ubuntuone.com/p/uB9/

ajlyon · May 21, 2011

The suggested change doesn't work for me either.

And the line endings are normalized within the translator to "\n" anyway:

if(tag == "N1" || tag == "N2" || tag == "AB" || tag == "KW") {
    // preserve line endings for N1/N2/AB fields, for EndNote
    // compatibility
    data += "\n"+rawLine;
}

No clue on this yet... But I think does usually work for people.