Using Bridge to Search for all Metadata containing a Phrase

[Jack]

I have been using Adobe Bridge to search a collection of images for those in which all metadata contain a phrase, with a degree of success. I knew I had loaded my keywords with an encyclopedic set of catch-all words, that were eating-up the search engine efforts. I was getting literally thousands of false positives. To fix this, I selected those false positives in Adobe Bridge, and deleted all keywords from them. Well and good, I thought, but no: many of those images are STILL returning a false positive, despite the fact that those keywords have been stripped out of them.

Question for you: where else in the image files, and using what recommended tool(s), shall I look to find these metadata in order to purge them?

[Stephen Marsh]

Open a file with a suspected “false positive” into Photoshop, then use the file menu/file info command and take a look at the appropriate metadata area for the keyword, such as Basic or Raw Data. This file info data can also be accessed from Bridge.

It could be that your search criteria in Bridge is incorrect.

For another opinion, you can use ExifTool to inspect the metadata.

The following command will test a top level/parent folder and all sub-level/child folders for a specific keyword titled “my-keyword”

exiftool -if '$subject =~ /my-keyword/' -subject -r '/Users/currentuser/Desktop/my folder'

If you simply wish to inspect a root folder/subfolders and list all of the keywords, then:

exiftool -subject -r '/Users/currentuser/Desktop/my folder'

To write a .CSV spreadsheet file of the files and their keywords, then:

exiftool -r -subject -csv '/Users/currentuser/Desktop/input folder' > '/Users/currentuser/Desktop/keyword-output.csv'

NOTE: These code examples are formatted for the Mac OS, MS Windows would use straight double quote marks " and the expected \ file path formatting.

[Jack]

Thanks much; very helpful. Found a false entry under Advanced | IPtC Core. That is where I will now do global edits to purge those entries.

[Jack]

Well, looking further at IPTC Core properties, I'm finding that to be a hit-or-miss proposition. After opening multiple files in Bridge, selecting File Info | IPTC Core, the field containing my desired "hot" entry,as found in the first example, did not even appear. Searched several other, single, "false-positive" files and did not find any entry containing the "false positive" value. So now I'm again on the lookout for a better tool to help me with this.

[Jack]

If I understand the exif tool description correctly, it will only _identify_ the image files that contain the "false positive" keyword, not the location of the offending phrase within photo metadata, which is a warren of (sub)sections,; searching through it is a time-consuming search process. I have, literally, hundreds of images currently returning false-positives on my system. I already know which files appear to contain the false positive; I just cannot localize it within the file(s), in order to purge it.

[Stephen Marsh]

You can use the following command to list ALL metadata in a file, which would return a lot more info than you may require, however if you have a needle in a haystack, you may need to look through the entire haystack!

exiftool -a -G1 -s -r 'path/to/folder/here'

Which is probably easier dealt with in a spreadsheet using find/search than in the command line interface output:

exiftool -a -G1 -s -r -csv 'path/to/folder/here' > '/Users/currentuser/Desktop/keyword-output.csv'

All of these ExifTool commands have been reading the metadata in the files, however one could also write/remove metadata instead. For example, ExifTool could be used to wipe out specific metadata entries in a folder/subfolders of files. So if you knew what fields you wished to wipe out it would be easy to do so.