Deleted keywords reappearing

Started by MikeB, May 18, 2017, 04:49:57 AM

Previous topic - Next topic

MikeB

I have been trying to delete some keywords from some jpeg file metadata in IMatch and they appear to be deleted, but after I run the Metadata writeback to update the jpegs, all the keywords are back. 
How do I PERMANENTLY delete the keywords?

Mario

This is usually caused by some non-synchronized keywords in IPTC/XMP contained in the image, in conjunction with the keyword import/mapping settings you have enabled and your thesaurus.

Sometimes this happens if the file contains IPTC keywords, but no standard IPTC version/id. IMatch then considers the file as not having IPTC data and does not update the existing partial IPTC record. The keywords ExifTool imports from the legacy IPTC fragment then override the XMP keywords on re-import after write-back.

You can check in the ExifTool Command Processor (Tools menu).
Run the "List Metadata" preset on one of your problem files.
Does it have IPTC keywords but no IPTC version number?
(You can attach the output of the ECP here so I can have a look).

To check if the legacy IPTC is the problem, do this:

1. Make a copy of one of your problem files (in IMatch, Ctrl+C, Ctrl+V).
2. Open the ExifTool Command Processor after selecting the copy.
3. Run the "Delete IPTC data" preset.

Now, add some keywords in the Keyword Panel and write-back.
If the keywords are now saved correctly, the IPTC was the problem.

Please attach/upload your file somewhere so we can have a look (or send it to my support email address, with a link to this thread).
-- Mario
IMatch Developer
Forum Administrator
http://www.photools.com  -  Contact & Support - Follow me on 𝕏 - Like photools.com on Facebook

jch2103

This updated information was very useful for my on-going efforts to clean up my older files. I had (still have...) a bunch of files with stubborn IPTC keywords that I had great trouble getting rid of. The above process works, although it's somewhat time-consuming due to the need for multiple metadata reads and writes to synchronize the tags.
John

Mario

Fixing corrupted, out-of-sync or missing metadata is always time-consuming. Fact.
It's one of the most expensive and complex tasks for every DAM 'migration' project. I had the pleasure to lend some expertise to some projects of that type and I could tell you stories... ;)

ExifTool has been designed to be safe first, fast second. It takes no shortcuts and always aims to produce the best and cleanest metadata stream. Which is perfectly fine for me!
When ExifTool rewrites IPTC data because of your changes in the ECP, IMatch has to reload the metadata because ExifTool will change digest data, timestamps, the XMP data cached in the IMatch database will no longer be current etc.

Reloading metadata will invalidate some (or many!) data-driven categories and collections. If IMatch needs this data to repaint the screen (File Window!) it has to update the data immediately and cannot wait for background processes. Also, if you use file versioning, your changes may require IMatch to update the data for multiple files, repeating the above steps for each file. Phew!

Tip:  You can search the IMatch log file for lines containing the #sl tag. I use this to highlight operations which take longer than 5 seconds. In many cases this is normal (e.g., loading a database) but it can pinpoint operations which take too long. To many automatic data-driven categories. Disk not keeping up, virus checker interfering by constantly scanning the database, things like that.
-- Mario
IMatch Developer
Forum Administrator
http://www.photools.com  -  Contact & Support - Follow me on 𝕏 - Like photools.com on Facebook