How to clear all Identifiers (Ids) except ISBN

[paul55391]

I want to just retain ISBNs for the books in my library. Is there a way to "bulk remove" every Identifier except existing ISBNs?

[Sabardeyn]

The only thing that I can think of is:

1. Open Tag Browser to determine all the identifiers used
2. Select one identifier at a time to limit affected books
3. Highlight all
4. Edit Metadata in Bulk
5. Search and Replace Tab
6. Search Mode: Regex
7. Search Field: identifiers
8. Search For: [expression, see below]
9. Replace With: empty
10. Verify test area results are correct
11. Apply

I cannot think of the exact regex you would need. Roughly, it would probably be the identifier's name, a colon and then "all characters, greedy, until the next whitespace/word end". So something like... ID:+.? Or is it ID:.+? Hmm.. think I'm missing the word boundary. Anyway, I'm not sure if the expression is written correctly. But hopefully you've got the idea. I imagine that the colon might need to be escaped.

I would definitely suggest trying this on a sample database. Just to make sure that it doesn't harm anything else. Maybe create a new library and then copy various sample books into it, run the regex and see what happens.

[eschwartz]

1. Select all the books you wish to affect, and open Bulk Metadata Edit. Make sure you are using regex mode.
2. Search for...
   

   Search field: identifiers
   Identifier type: isbn
   Search for: (.*)
   Replace with: \1
   Destination field: <anything you can spare. I used "comments">

   Apply the edit.
3. Search for...
   

   Search field: comments
   Search for: (.*)
   Replace with: isbn:\1
   Destination field: identifiers
   Identifier type: *

   Apply the edit.

---

The reason this works is because, while there is no way to replace an identifier with a subset of itself, you can use * to replace all the identifiers IF AND ONLY IF the search field is not "identifiers". So by temporarily storing the isbn in another field, you can have the isbn overwrite the whole identifiers field.
(I used the comments field, because I used a test book created by importing another book from the same library. )

See screenshots.

[eschwartz]

Originally Posted by Sabardeyn

I cannot think of the exact regex you would need. Roughly, it would probably be the identifier's name, a colon and then "all characters, greedy, until the next whitespace/word end". So something like... ID:+.? Or is it ID:.+? Hmm.. think I'm missing the word boundary. Anyway, I'm not sure if the expression is written correctly. But hopefully you've got the idea. I imagine that the colon might need to be escaped.

When S&Ring the identifiers field, each identifier acts as a self-contained unit. So the regex would be ".*", and replaced with a null value.

I would definitely suggest trying this on a sample database. Just to make sure that it doesn't harm anything else. Maybe create a new library and then copy various sample books into it, run the regex and see what happens.

I second the idea of testing this out first, always a good idea when doing Bulk S&R -- with a powerful tool like that, you have room to mess up quite badly.

[chaley]

You can also use the technique described in this thread: https://www.mobileread.com/forums/sho...d.php?t=207433

[eschwartz]

Originally Posted by chaley

You can also use the technique described in this thread: https://www.mobileread.com/forums/sho...d.php?t=207433

Did not realize you can use a template as the search field. That is just really cool.

[eschwartz]

... And yes, apparently I actually answered a question in that thread, and still didn't notice.

[chaley]

Originally Posted by eschwartz

Did not realize you can use a template as the search field. That is just really cool.

I added that feature to allow more-or-less arbitrary field values such as setting destination fields to values constructed from other fields. Turns out to have many other uses, including this one (pruning values in multiple fields). It isn't the easiest feature to wrap one's head around, I admit.

[paul55391]

eschwartz, I learned something from your post (#3), but your method wipes out 'comments', and I already use that field. In fact, there isn't one single field I can unconditionally wipe out. Maybe it's possible to first create a temporary field to store the isbn value (and then throw the temporary field away.) Does this sound like a reasonable approach?

[BetterRed]

@paul55391 - IMO that would work, its easy enough to add and delete custom columns and there's a comments like custom column type.

Custom columns are always referenced with a '#' in front of the internal name - so if yours was 'temporary' you would use #temporary instead of comments in eschwartz's post #3.

BR

[eschwartz]

A better way would be to use chaley's method, which I would've recommended if I knew it existed.

Search field: {template}         (YES! This exists and it is awesome!)
Template: {isbn:|isbn:|}
Search for: (.*)
Replace with: isbn:\1
Destination field: identifiers
	>> Mode: Replace field, Split Result="checked"
Identifier type: *

Template uses isbn # and prepends "isbn:" to it, then feeds that into the replace. No need for a custom column at all. Thanks chaley!!

[paul55391]

Thanks, that works fine. (I did change the "Replace with" so it agrees with your screen snap:
Replace with: \n
Without this change the string "isbn:" is duplicated.

Thanks a lot. Now that I understand what one can do with regular expressions, I can see this might be very useful in the future as well.