Regex help for importing books from filename

CoreyEMTP · 02-09-2026, 03:09 PM

Hi All,

I'm hoping a calibre regex guru can help me out with this. I have failed, and even AI couldn't help me. Here's a filename example:

A. J. Author, Author Two - This Is the Title (2020) [1234567890123].epub

The published date and isbn are independently optional and potentially not in the filename. Here's the closest i got, which only works if the optional items are present:

(?P<author>.*?) - (?P<title>.*?) \((?P<published>\d{4})\) \[(?P<isbn>.*?)\]

The closest I got with AI was one where it would work if both optional items are present, or if only the isbn. if there is neither or only the year, it returns the entire filename minus extension as the title with no other information other than "unknown" for the author(s).

Any help for a regex that works for all four cases would be greatly appreciated!

mitsie · 02-11-2026, 04:15 PM

Why would you need to use regex? You said the AI couldn't help you, but that was because you asked it to provide you with a straight RegEx to try to catch all. Use the AI to read the book and pull the author(s) and title out for you, with the temperature turned to 0, and you will get the answer correct every time..

It saves messing around with Regex and makes life easier

BetterRed · 02-11-2026, 05:03 PM

Quote:

Originally Posted by CoreyEMTP

. . . even AI couldn't help me.

If course it couldn't. AI should be regarded is as being: dumb as an ox, thick as two short planks or a brick, witless as a whelk… what ever.

Besides which there's more than one RegEx, I forget which one calibre uses.

BR

mitsie · 02-11-2026, 05:37 PM

Quote:

Originally Posted by BetterRed

AI should be regarded is as being: dumb as an ox, thick as two short planks or a brick, witless as a whelk… what ever.

True, hitting a screw with the back of a power drill isn't the correct way to use that tool, but it might get the job done in a sloppy and inaccurate fashion.. It's the people who learn to turn the power drill around and use the trigger that get better results from the tool..

DNSB · 02-11-2026, 06:36 PM

Quote:

Originally Posted by BetterRed

Besides which there's more than one RegEx, I forget which one calibre uses.
BR

If I remember correctly, calibre uses Python's re regex library.

When I use regex101.com, for calibre, I select Python and for Sigil, I select PCRE2.

CoreyEMTP · 02-12-2026, 12:31 AM

I appreciate the suggestions, and I did attempt to formulate a regex myself, but apparently there are some features of python regex that are not found in calibre's implementation. I'll run some through the regex helper, but I'm not confident.

CoreyEMTP · 02-12-2026, 12:34 AM

Well, I came up with this one that works perfectly fine on regex101, using the python engine:

(?P<author>[^-]+) - (?P<title>[^(\[]+)(?: \((?P<year>\d{4})\))?(?: \[(?P<isbn>[^\]]+)\])?

Doesn't work in calibre however.

theducks · 02-12-2026, 02:45 PM

Calibre has no field named Year or ISBN
published?
identifiers:isbn?

CoreyEMTP · 02-13-2026, 05:48 PM

You're absolutely right about the year, I don't know how I overlooked that. It does have a (?P<isbn>) in Calibre, however. Thanks!

theducks · 02-13-2026, 06:07 PM

Quote:

Originally Posted by CoreyEMTP

It does have a (?P<isbn>) in Calibre, however. Thanks!

Not any longer in Calibre 9

CoreyEMTP · 02-13-2026, 06:17 PM

For those curious or who have a similar file layout, here's the regex that finally worked:

Code:

^(?P<author>.+?)\s-\s(?P<title>(?:(?!\s[\(\[]).)+)(?:\s\((?P<published>\d{4})\))?(?:\s\[(?P<isbn>\d{10,13})\])?$

...and this is in 9.2.1, so yes it absolutely does have a (?P<isbn>)

theducks · 02-13-2026, 07:19 PM

Quote:

Originally Posted by CoreyEMTP

For those curious or who have a similar file layout, here's the regex that finally worked:

Code:

^(?P<author>.+?)\s-\s(?P<title>(?:(?!\s[\(\[]).)+)(?:\s\((?P<published>\d{4})\))?(?:\s\[(?P<isbn>\d{10,13})\])?$

...and this is in 9.2.1, so yes it absolutely does have a (?P<isbn>)

Then Kovid must have made a special x-reference to identifiers, as He removed the (unused) ISBN column according to the release notes.

DNSB · 02-13-2026, 08:35 PM

AFAIR, ISBNs are part of the identifiers table now.

CoreyEMTP · 02-13-2026, 09:50 PM

That I don't know, but it does have the item available under the specified settings. The test box reads and displays the parts of the filename in their respective fields, including ISBN.

theducks · 02-13-2026, 10:42 PM

test box?

Do you mean in the MDE or details section?

02-09-2026, 03:09 PM	#1
CoreyEMTP Junior Member Posts: 9 Karma: 10 Join Date: Feb 2026 Device: Kobo	Regex help for importing books from filename Hi All, I'm hoping a calibre regex guru can help me out with this. I have failed, and even AI couldn't help me. Here's a filename example: A. J. Author, Author Two - This Is the Title (2020) [1234567890123].epub The published date and isbn are independently optional and potentially not in the filename. Here's the closest i got, which only works if the optional items are present: (?P<author>.?) - (?P<title>.?) \((?P<published>\d{4})\) \[(?P<isbn>.*?)\] The closest I got with AI was one where it would work if both optional items are present, or if only the isbn. if there is neither or only the year, it returns the entire filename minus extension as the title with no other information other than "unknown" for the author(s). Any help for a regex that works for all four cases would be greatly appreciated!

02-11-2026, 04:15 PM	#2
mitsie Member Posts: 23 Karma: 12814 Join Date: Jun 2025 Device: PocketBook 4	Why would you need to use regex? You said the AI couldn't help you, but that was because you asked it to provide you with a straight RegEx to try to catch all. Use the AI to read the book and pull the author(s) and title out for you, with the temperature turned to 0, and you will get the answer correct every time.. It saves messing around with Regex and makes life easier Last edited by mitsie; 02-11-2026 at 04:17 PM.

02-13-2026, 05:48 PM	#9
CoreyEMTP Junior Member Posts: 9 Karma: 10 Join Date: Feb 2026 Device: Kobo	You're absolutely right about the year, I don't know how I overlooked that. It does have a (?P<isbn>) in Calibre, however. Thanks! Last edited by CoreyEMTP; 02-13-2026 at 05:50 PM.

02-13-2026, 06:17 PM	#11
CoreyEMTP Junior Member Posts: 9 Karma: 10 Join Date: Feb 2026 Device: Kobo	For those curious or who have a similar file layout, here's the regex that finally worked: Code: ^(?P<author>.+?)\s-\s(?P<title>(?:(?!\s[\(\[]).)+)(?:\s\((?P<published>\d{4})\))?(?:\s\[(?P<isbn>\d{10,13})\])?$ ...and this is in 9.2.1, so yes it absolutely does have a (?P<isbn>) Last edited by CoreyEMTP; 02-13-2026 at 06:19 PM. Reason: Wrong layout

02-13-2026, 10:42 PM	#15
theducks Well trained by Cats Posts: 31,512 Karma: 62503986 Join Date: Aug 2009 Location: The Central Coast of California Device: Kobo Libra2,Kobo Aura2v1, K4NT(Fixed: New Bat.), Galaxy Tab A	test box? Do you mean in the MDE or details section? Attached Thumbnails

02-12-2026, 12:31 AM	#6
CoreyEMTP Junior Member Posts: 9 Karma: 10 Join Date: Feb 2026 Device: Kobo	I appreciate the suggestions, and I did attempt to formulate a regex myself, but apparently there are some features of python regex that are not found in calibre's implementation. I'll run some through the regex helper, but I'm not confident.

02-12-2026, 12:34 AM	#7
CoreyEMTP Junior Member Posts: 9 Karma: 10 Join Date: Feb 2026 Device: Kobo	Well, I came up with this one that works perfectly fine on regex101, using the python engine: (?P<author>[^-]+) - (?P<title>[^(\[]+)(?: \((?P<year>\d{4})\))?(?: \[(?P<isbn>[^\]]+)\])? Doesn't work in calibre however.

02-12-2026, 02:45 PM	#8
theducks Well trained by Cats Posts: 31,512 Karma: 62503986 Join Date: Aug 2009 Location: The Central Coast of California Device: Kobo Libra2,Kobo Aura2v1, K4NT(Fixed: New Bat.), Galaxy Tab A	Calibre has no field named Year or ISBN published? identifiers:isbn?

02-13-2026, 08:35 PM	#13
DNSB Bibliophagist Posts: 50,294 Karma: 178402650 Join Date: Jul 2010 Location: Vancouver Device: Kobo Sage, Libra Colour, Lenovo M8 FHD, Paperwhite 4, Tolino epos	AFAIR, ISBNs are part of the identifiers table now.

02-13-2026, 09:50 PM	#14
CoreyEMTP Junior Member Posts: 9 Karma: 10 Join Date: Feb 2026 Device: Kobo	That I don't know, but it does have the item available under the specified settings. The test box reads and displays the parts of the filename in their respective fields, including ISBN.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
regex for importing books with comment	Widukind	Library Management	5	02-12-2021 06:13 AM
Adding Books - Filename Regex Management	Paulie_D	Library Management	4	05-10-2015 02:03 PM
Adding books - filename RegEx author FN (initials) LN	Rob557	Library Management	8	10-16-2014 05:53 PM
regex for filename with muliple dots? (adding books)	kite	Library Management	5	09-28-2014 10:09 AM
Need a regex for importing books	flinkdeldinky	Calibre	31	10-29-2011 09:31 AM