How to clean up junk indentifier types?

[capnm]

I've been ignoring all the junk identifier types that have appeared in the identifiers field long enough -- so now I want to get rid of a few hundred of them (mostly ISBNs but other cruft too).

I don't see any easy way to do it though.

I was thinking the most likely would be to use the Bulk Edit Search & Replace function to back up the few I want to keep in spare columns, then just blank the whole identifiers field.

But the only way I see to blank the whole field is one book at a time via Edit Metadata.

Alternatively, removing the identifier types one at a time via Search and Replace is too tedious to contemplate.

Is there another way I've overlooked?


(Am I really the only person with this problem? The cruft comes from mainstream retail epubs, so I find that hard to believe.)

[kiwidude]

@capnm - yeah I do the same thing for the same reasons, and yeah they are a pain to remove. It has you cursing the teeny tiny identifiers textbox in the Edit Metadata dialog when you edit them one by one and try to find the right place in the very long line to chop them out of

I tend to use the tag browser, click on them one at a time to bring up the matching set of books for that identifier and then do the edit as you say.

I know there is limited space on that dialog, but they are the sort of data that you want in a list to easily click on an X to delete etc, dunno how it would fit in though without making it a popup.

[chaley]

I submitted code changes to permit arbitrary changes to the identifiers field in bulk search/replace. If the source field is not "identifiers" and the destination field is "identifiers" and the destination identifier is "*", then the result of the search/replace is used to replace the entire identifiers field. To empty the field, use "{template}" as the source field, ".*" as the search for string, "identifiers" as the destination field, and "*" as the destination identifier.

Note that you can construct any set of identifiers you wish in the search/replace operation, making it possible to remove all but some identifiers.

Note also that the string used to generate the resulting identifiers field must be a (possibly null) comma-separated set of colon-separated pairs of strings.

[capnm]

That's slick, thanks!

So to keep just isbn & goodreads, is this correct?

Search field: {template}
Template: {isbn:|isbn:|, }{identifiers:select(goodreads)|goodreads:|}
Search for: empty
Replace with: empty
Destination field: identifiers
Identifier type:*
Mode: Replace field
Split Result: Unchecked

Do I care about this possibly creating an extraneous trailing comma?

[chaley]

Originally Posted by capnm

That's slick, thanks!

So to keep just isbn & goodreads, is this correct?

Search field: {template}
Template: {isbn:|isbn:|, }{identifiers:select(goodreads)|goodreads:|}
Search for: empty
Replace with: empty
Destination field: identifiers
Identifier type:*
Mode: Replace field
Split Result: Unchecked

Yes, except you should leave Split Result checked so that each comma-separated item is treated as a separate identifier. If you uncheck the box then SR tries to treat the string as 1 identifier and will complain that it finds too many colons.

Do I care about this possibly creating an extraneous trailing comma?

No, as long as there aren't any non-blank characters following it. Null items are ignored, which means that ",,,,,,,,isbn:12345,,,,,goodreads:81267,,,,,," is also OK.

[ssholloway]

Many thanks for this. I did discover two minor changes to the formula above that are required for Calibre 1.39. I wanted to keep the ASIN so here's what worked for me:

Search field: {template}
Template: {isbn:|isbn:|, }{identifiers:select(amazon)|amazon:|}
Search for: {template}
Replace with: empty
Destination field: identifiers
Identifier type:*
Mode: Replace field
Split Result: checked
With values separated by:, (i.e., a comma followed by a space)

[Tatjana]

It is possible at this time, using latest Version of Calibre for Windows (1.39) get rid of Identifiers that do not have a identifier prefix?

example: isbn:9780310231202, 0310231205:0310231205

Is it possiible to get rid of all identifiers that do not have a prefix like "isbn:", "amazon:" etc?
Please note the example above has a "junk prefix "0310231205:"

If it is not possible to this on Calibre right now, is there any free sqlite database editor what would allow me to do this?

Thank you very much in advance for any help.

[eschwartz]

Just edit it normally. There is no way for calibre to know what is a "junk identifier".

Copying it into a text editor or Word should help you see what you are dealing with a little better than scrolling through the tiny edit box.

[chaley]

Originally Posted by Tatjana

It is possible at this time, using latest Version of Calibre for Windows (1.39) get rid of Identifiers that do not have a identifier prefix?

example: isbn:9780310231202, 0310231205:0310231205

Is it possiible to get rid of all identifiers that do not have a prefix like "isbn:", "amazon:" etc?
Please note the example above has a "junk prefix "0310231205:"

If it is not possible to this on Calibre right now, is there any free sqlite database editor what would allow me to do this?

Thank you very much in advance for any help.

Probably the best you can do in calibre, and also I suspect in an SQL db utility, is to keep identifiers if they are in a good identifiers list. However, if you can figure out a regexp (or something) that is guaranteed to match a junk identifier but not a good one then you could do that.

The "good list" technique would use the template approach above to select out the ones you wish to keep.

[Tatjana]

Originally Posted by chaley

Probably the best you can do in calibre, and also I suspect in an SQL db utility, is to keep identifiers if they are in a good identifiers list. However, if you can figure out a regexp (or something) that is guaranteed to match a junk identifier but not a good one then you could do that.

The "good list" technique would use the template approach above to select out the ones you wish to keep.

Excellent post (all your posts on this thread actually.)

You have helped me a whole lot!

a BIG! Thank you!