[Regex Search] Minimal match not possible?

nqk · 12-22-2014, 10:03 PM

Dear you guys,

When I do a regex search for something, just say “(.*)”, Editor would select EVERY thing between the first “ and the last ” in a paragraph, if that paragraph includes multiple set of “(.*)”. This means that the selection is WRONG, because it includes texts in between two sets of “(.*)”.

If the box "Dot all" is checked, the situation applies then not for a single paragraph but the whole file.

In Sigil, there is a "Minimal Match" option, which would select the right set.

kovidgoyal · 12-22-2014, 10:08 PM

(.*?)

dmonasse · 12-23-2014, 08:28 AM

Look at the Python module "re" documentation at https://docs.python.org/2/howto/regex.html#regex-howto
and you shall read:
the solution is to use the non-greedy qualifiers *?, +?, ??, or {m,n}?, which match as little text as possible.

eschwartz · 12-23-2014, 12:14 PM

Possibly a good idea to get used to doing this in the regex itself on general principle, rather than associating instruction sets with the regex.

DigitEditLab · 12-23-2014, 04:34 PM

Additional request, @kovid or any expert at hand:

(.*?) is a bit hungry sometimes. I'd like to match anything but the tags: only characters, digits, punctuation marks and spaces but no <>.

What's the trick? Thanks in advance.

eschwartz · 12-23-2014, 04:40 PM

The trick is don't use a dot-match-all symbol. Use a regex character class, like

Code:

[a-zA-Z0-9.?',"]

Or match all but tag brackets:

Code:

[^<>]

I found this regex tutorial to be very helpful in learning the various fine points of regex: http://www.regular-expressions.info/

There are some very interesting yet obscure applications in the corners.

Like this interesting use of negative lookarounds to find matching span tags, even when nested, and delete the matching sets:

Code:

<span[^<>]*>((?:(?!<(?:/?span)).)*)</span>

The bit on the inside finds only text that does not include the arbitrary string "</?span", inside matching span tags.

mikapanja · 12-24-2014, 01:56 AM

Quote:

Originally Posted by DigitEditLab

Additional request, @kovid or any expert at hand:

(.*?) is a bit hungry sometimes. I'd like to match anything but the tags: only characters, digits, punctuation marks and spaces but no <>.

Just in case you need to match everything between the tags, excluding the tags, this should do the trick:

(?<=<.*?>)(.*)(?=<.*?>)

nqk · 12-24-2014, 03:19 AM

Thank you, guys.

Each day I learn something.

12-22-2014, 10:03 PM	#1
nqk Fanatic Posts: 564 Karma: 32228 Join Date: Feb 2012 Device: Onyx Boox Leaf	[Regex Search] Minimal match not possible? Dear you guys, When I do a regex search for something, just say *“(.)”, Editor would select EVERY thing between the first “ and the last ” in a paragraph, if that paragraph includes multiple set of “(.)”. This means that the selection is WRONG, because it includes texts in between two sets of “(.)”**. If the box "Dot all" is checked, the situation applies then not for a single paragraph but the whole file. In Sigil, there is a "Minimal Match" option, which would select the right set.

12-23-2014, 04:40 PM	#6
eschwartz Ex-Helpdesk Junkie Posts: 19,421 Karma: 85400180 Join Date: Nov 2012 Location: The Beaten Path, USA, Roundworld, This Side of Infinity Device: Kindle Touch fw5.3.7 (Wifi only)	The trick is don't use a dot-match-all symbol. Use a regex character class, like Code: [a-zA-Z0-9.?',"] Or match all but tag brackets: Code: [^<>] I found this regex tutorial to be very helpful in learning the various fine points of regex: http://www.regular-expressions.info/ There are some very interesting yet obscure applications in the corners. Like this interesting use of negative lookarounds to find matching span tags, even when nested, and delete the matching sets: Code: <span[^<>]>((?:(?!<(?:/?span)).))</span> The bit on the inside finds only text that does not include the arbitrary string "</?span", inside matching span tags. Last edited by eschwartz; 12-23-2014 at 04:51 PM.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Search and replace returns entire field when no match is found	wladdy	Calibre	2	02-16-2014 01:51 AM
Metadata Search & Replace - when it doesn't match	Aldebaranian	Library Management	4	09-28-2011 11:35 AM
how to have regex dot match any character including newline?	gnychis	Calibre	5	11-30-2010 06:35 PM
Need help with a conversion regex - can't match newline	ereader123	Calibre	2	03-29-2010 10:58 AM
Search tags using exact match?	chaley	Calibre	3	01-21-2010 01:16 PM

12-22-2014, 10:08 PM	#2
kovidgoyal creator of calibre Posts: 45,265 Karma: 27110894 Join Date: Oct 2006 Location: Mumbai, India Device: Various	(.*?)

12-23-2014, 08:28 AM	#3
dmonasse Member Posts: 23 Karma: 10 Join Date: Apr 2014 Location: Paris Device: ipad 2, Ubuntu	Look at the Python module "re" documentation at https://docs.python.org/2/howto/regex.html#regex-howto and you shall read: the solution is to use the non-greedy qualifiers ?, +?, ??, or {m,n}?, which match as little text as possible.*

12-23-2014, 12:14 PM	#4
eschwartz Ex-Helpdesk Junkie Posts: 19,421 Karma: 85400180 Join Date: Nov 2012 Location: The Beaten Path, USA, Roundworld, This Side of Infinity Device: Kindle Touch fw5.3.7 (Wifi only)	Possibly a good idea to get used to doing this in the regex itself on general principle, rather than associating instruction sets with the regex.

12-23-2014, 04:34 PM	#5
DigitEditLab Member Posts: 21 Karma: 10 Join Date: Feb 2014 Device: Kobo Aura HD, Samsung Note II, Kindle and a few more	Additional request, @kovid or any expert at hand: (.?) is a bit hungry sometimes. I'd like to match anything but the tags*: only characters, digits, punctuation marks and spaces but no <>. What's the trick? Thanks in advance.

12-24-2014, 03:19 AM	#8
nqk Fanatic Posts: 564 Karma: 32228 Join Date: Feb 2012 Device: Onyx Boox Leaf	Thank you, guys. Each day I learn something.

Advert

Advert