Illegal Characters in Keywords

Report · Nov 16, 2017

Could anybody offer an explanation as to why Bridge is perfectly happy to allow the use of the characters < and > in keywords, but not Lightroom?

Unfortuneatly, I have used these characters a lot in Bridge over many years for my own good reasons.

When such images are imported into Lightroom, these characters are stripped out as they are deemed to be illegal characters.

This becomes a problem if metadata is subsequently saved from Lightroom back to the original images as these characters will have disappeared, thus rather messing up my system.

Report · Nov 16, 2017

Unfortunately, questions requesting an explanation for why something does something almost always go unanswered. Bridge was released in 2005 and Lightroom in 2007 - both more than ten years ago. The reasons for why one works one way and the other, if originally deliberate may be lost in the depths of time. It may originate from the programs' purpose for they were created for different audiences.

Your best option is to request a feature change for Lightroom via this site: Lightroom Classic CC | Photoshop Family Customer Community

Report · Nov 16, 2017

I observe different behavior. In Bridge 2018, I assigned the keyword "MMM>NNN" to a photo:

I then imported that photo into LR 7.0.1. It correctly imported the keyword as written:

Next, in LR I did Metadata > Save Metadata To File, and observed the same keyword still correctly assigned in Bridge. ExifTool also shows that LR saved the keywords correctly to the metadata fields:

$ exiftool -xmp:subject -xmp:hierarchicalsubject -iptc:keywords p.jpg

Subject : MMM>NNN

Hierarchical Subject : MMM>NNN

Keywords : MMM>NNN

To find out what's going on, please report the exact versions you're running and Windows / Mac. in Bridge, please do Help > About Bridge (Windows) or Adobe Bridge > About Bridge (Mac). In LR, do Help > System Info.

Report · Nov 16, 2017

Thanks for your reply.

Bridge and Lightroom are running on a Windows 10 PC.

Bridge Version:

Lightroom System Info:

Here is a single example of keywords applied to an image in Bridge with < and > characters:

and here are the same keywords after the image was imported into Lightroom:

As you can see the < and > characters have been stripped and then the keywords are listed alphabetically though clearly in a different order.

Report · Nov 16, 2017

Looking at this, your keywords with less than/greater than are being treated hierarchically as if this was a separator key. See the Lightroom Preferences:

I would call that a bug as those symbols, while restricted from file renaming are not listed as keyword demarcation characters. I would file it.

Report · Nov 16, 2017

It's definitely a bug in LR, but it depends on undocumented/buggy behavior in Bridge as well. If you add the keywords in Bridge using the Keywords panel, the bug doesn't occur:

But if you add the keywords using the IPTC (IIM, legacy) panel, the bug occurs:

When you add the keywords using the Keywords panel, Bridge records the keyword in all three keyword metadata fields required by the Metadata Working Group spec (to which Adobe subscribes):

 $ exiftool -a -G -xmp:subject -xmp:hierarchicalsubject -iptc:keywords p.jpg
[XMP]           Subject                         : <B>
[XMP]           Hierarchical Subject            : <B>
[IPTC]          Keywords                        : <B>

But when you add the keywords using the IPTC (IIM, legacy) panel, Bridge only adds the keywords to two of those fields:

$ exiftool -a -G -xmp:subject -xmp:hierarchicalsubject -iptc:keywords p.jpg
[XMP]           Subject                         : <B>
[IPTC]          Keywords                        : <B>

This isn't documented in the Bridge manual, and it doesn't follow the MWG spec.

It appears that when XMP:HierarchicalSubject is present, LR correctly interprets "<" and ">" in the keywords. But when it isn't present, it incorrectly handles those characters.

Report · Nov 17, 2017

Yes, I see that, in Lightroom preferences, my < and > symbols are not permitted in file names, but should be allowed in keywords.

How do I report this issue as a bug please?

Interestingly, I just discovered that I can either add < and > to the keywords in Lightroom afterwards or edit the keyword tags to include the characters required. It is the import procedure which strips them.

Report · Nov 17, 2017

The site to which I linked for the feature request is the proper place to report the bug. I think John Ellis has a thread going there already. I would recommend adding your information to John's thread. Once posted there, a Customer Support Representative will test the behavior and verify it, and file the bug in Adobe's bug system.

Here is John's thread: https://feedback.photoshop.com/photoshop_family/topics/lightroom-read-metadata-from-file-incorrectly...

Lightroom: Read Metadata From File incorrectly handles ">" and [less than] in keywords | Photoshop F...

Add your information there.

I will give it a plus 1 as soon as I see you add your info.

Report · Nov 17, 2017

Just to correct the record for future reference: Contrary to what I wrote above, the Metadata Working Group spec doesn't say anything about XMP:HierarchicalSubject, which is an undocumented Adobe-specific field.

Report · Feb 02, 2020

Thanks for all of your input into this but I haven't heard a word further about this issue in two years. Is there anything to tell?