Regex help needed.

patrik · 01-27-2019, 02:19 PM

I have plenty of similar to the following:

<p nodeindex="67">There is a great episode of Red Dwarf, where the crew fall prey to the “Despair Squid” on an ocean planet while salvaging from a derelict vessel at the bottom of the ocean. The squid squirted the ship with ink that had hallucinogenic properties that created despair in the crew members. They realise that they have been infected and head back to their own ship incredibly quickly to get a mood stabiliser. On the lift back to the top of the ship Dave Lister says “<em nodeindex="318">I don’t seem to have been effected…. Though it is true to say that no-one ever truly loved me in my entire life…</em>.” and starts crying as despair sets in. It is really funny!</p>

... and want to remove all nodeindex="xx" entries and I'm trying and googling all I can about regex.... I usually can figure out things like this after a while but...

Something like nodeindex=".*" matches way to much.

(I'm using Sigil.)

jhowell · 01-27-2019, 02:34 PM

.* is greedy. It will match as many characters as possible.
.*? will match as few characters as possible and will work better here.
[0-9]+ is another alternative if the value is always numeric.

patrik · 01-27-2019, 02:58 PM

Ah, nodeindex="[0-9]+" worked perfectly!

Thank you very much!!

The regex-sites I find are all very complex. Isn't there some Regex for dummies out there?

elchamaco · 01-28-2019, 05:01 AM

You can use also use \d+ to find any number \d for numbers and the + for following numbers

<p nodeindex="\d+">

and replace with <p>

patrik · 01-28-2019, 05:16 AM

Ok, thank you.

patrik · 01-28-2019, 08:13 AM

If I may...

I want to remove a beginning div.

It's like this:

<body>
<div lang="en">

(bunch of stuff)

</div>
</body>

I thought something like <div lang="en">.*</div> would match.

DNSB · 01-28-2019, 12:44 PM

Quote:

Originally Posted by patrik

If I may...

I want to remove a beginning div.

It's like this:

<body>
<div lang="en">

(bunch of stuff)

</div>
</body>

I thought something like <div lang="en">.*</div> would match.

You might want to make sure that you match the last </div> in the file. So perhaps <body> <div lang="en">(.*?)</div> </body>
and replace with
<body>\1</body>

As usual, make sure you have a backup before making changes.

patrik · 01-28-2019, 01:15 PM

Thanks, that's along what I thought would work, but it doesn't match anything.

This is a bit frustrating....

DaleDe · 01-28-2019, 01:45 PM

Quote:

Originally Posted by patrik

Ah, nodeindex="[0-9]+" worked perfectly!

Thank you very much!!

The regex-sites I find are all very complex. Isn't there some Regex for dummies out there?

You could check our wiki Regular expressions

Dale

patrik · 01-28-2019, 02:01 PM

Quote:

Originally Posted by DaleDe

You could check our wiki Regular expressions

Dale

Thankyou.

Reading that got me thinking... could it be that it not matching is due to new line not being included in '.*' ?

Ok, I tried with <div lang="en">.*|\n*</div> which actually seems to find/match.

But when I add ( ) to be able to use \1, it doesn't match anymore.

Am I doing this wrong: <div lang="en">(.*|\n*)</div> ?

patrik · 01-28-2019, 02:34 PM

Hey, enabling DotAll in Sigil made this work: <div lang="en">(.*)</div>

Do I understand correct that . now also includes new line and that's why it works?

DNSB · 01-28-2019, 05:40 PM

Quote:

Originally Posted by patrik

Thanks, that's along what I thought would work, but it doesn't match anything.

This is a bit frustrating....

Did you copy/paste the first & last bits? In your example, copy/pasting:

Code:

<body>
<div lang="en">

would keep the EOL between the > and <. Ditto for the:

Code:

</div>
</body>

The search line would show a space between the > and < where the EOL was. In the sample ebook I used for testing, the find showed as

Code:

<body> <div class="_override_2">(.*?)</div> </body>

with the replace showing as:

Code:

<body>\1</body>

I have DotAll, Auto-Tokenize and Wrap checked in the options.

Tex2002ans · 02-03-2019, 05:58 PM

Quote:

Originally Posted by patrik

The regex-sites I find are all very complex. Isn't there some Regex for dummies out there?

Regular-expressions.info is the go-to place for learning Regular Expressions.

"For dummies", the "Regex Examples" thread in Sigil's forums is good to see a lot of real life examples. Lots of people have asked questions over the years, and people have answered.

I've written multiple posts over the years breaking down Regex into bite-sized chunks, and I try to explain "in normal English" like:

Or sometimes color coding:

ALL CAPS to First Caps

But really, there's not much you can do besides learning the basics. You then just build them up piece-by-piece, and as long as you know what each little piece does, you can tackle the more complicated patterns you recognize.

Last year, I was training a few people in the basics of ebook creation. I created a presentation, and Regex was one of the sections (10 slides).

I showed a handful of common issues that creep up like:

How to find Page Numbers (p. 123, Page 234)
Finding 3 capital letters in a row (ABCs, FBI, XYZ)
Finding lowercase letter in beginning of paragraph
Correcting spaces around em dashes
Finding all the "1 a.m." and "9:32 p.m."
Finding roman numeral chapters (Chapter III)
Paragraphs that end with no punctuation
Finding/Correcting punctuation OUTSIDE of quotation marks
Replacing 98-99 (hyphen) or 98—99 (em dash) with the proper en dash
[...]

If anyone is interested in those slides, send me a PM. :P

DaleDe · 02-04-2019, 12:19 PM

Sounds like a good addition to our wiki.

Dale

patrik · 02-04-2019, 12:53 PM

Tex, this forum has some really exellent contributers, and you sure are one of them. Thank you!

Will send you a pm.

01-27-2019, 02:19 PM	#1
patrik Guru Posts: 682 Karma: 4568205 Join Date: Jan 2010 Location: Sweden Device: Kobo Forma	Regex help needed. I have plenty of similar to the following: <p nodeindex="67">There is a great episode of Red Dwarf, where the crew fall prey to the “Despair Squid” on an ocean planet while salvaging from a derelict vessel at the bottom of the ocean. The squid squirted the ship with ink that had hallucinogenic properties that created despair in the crew members. They realise that they have been infected and head back to their own ship incredibly quickly to get a mood stabiliser. On the lift back to the top of the ship Dave Lister says “<em nodeindex="318">I don’t seem to have been effected…. Though it is true to say that no-one ever truly loved me in my entire life…</em>.” and starts crying as despair sets in. It is really funny!</p> ... and want to remove all nodeindex="xx" entries and I'm trying and googling all I can about regex.... I usually can figure out things like this after a while but... Something like nodeindex=".*" matches way to much. (I'm using Sigil.)

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Regex Help Needed	gstock97	Library Management	1	09-01-2015 07:13 PM
Help with regex needed.	kamanza	Library Management	3	01-24-2012 07:27 AM
RegEx Help needed	ghostyjack	Sigil	14	11-02-2011 10:22 AM
RegEx REPLACEMENT: Help needed!	LARdT	Sigil	12	01-04-2010 07:25 PM
Regex help needed	gandor62	Calibre	2	11-04-2009 10:27 AM

01-27-2019, 02:34 PM	#2
jhowell Grand Sorcerer Posts: 7,059 Karma: 91577715 Join Date: Nov 2011 Location: Charlottesville, VA Device: Kindles	.* is greedy. It will match as many characters as possible. .*? will match as few characters as possible and will work better here. [0-9]+ is another alternative if the value is always numeric.

01-27-2019, 02:58 PM	#3
patrik Guru Posts: 682 Karma: 4568205 Join Date: Jan 2010 Location: Sweden Device: Kobo Forma	Ah, nodeindex="[0-9]+" worked perfectly! Thank you very much!! The regex-sites I find are all very complex. Isn't there some Regex for dummies out there?

01-28-2019, 05:01 AM	#4
elchamaco Zealot Posts: 128 Karma: 500 Join Date: Aug 2011 Device: kindle, boox	You can use also use \d+ to find any number \d for numbers and the + for following numbers <p nodeindex="\d+"> and replace with <p>

01-28-2019, 05:16 AM	#5
patrik Guru Posts: 682 Karma: 4568205 Join Date: Jan 2010 Location: Sweden Device: Kobo Forma	Ok, thank you.

01-28-2019, 08:13 AM	#6
patrik Guru Posts: 682 Karma: 4568205 Join Date: Jan 2010 Location: Sweden Device: Kobo Forma	If I may... I want to remove a beginning div. It's like this: <body> <div lang="en"> (bunch of stuff) </div> </body> I thought something like <div lang="en">.*</div> would match.

01-28-2019, 01:15 PM	#8
patrik Guru Posts: 682 Karma: 4568205 Join Date: Jan 2010 Location: Sweden Device: Kobo Forma	Thanks, that's along what I thought would work, but it doesn't match anything. This is a bit frustrating....

01-28-2019, 02:34 PM	#11
patrik Guru Posts: 682 Karma: 4568205 Join Date: Jan 2010 Location: Sweden Device: Kobo Forma	Hey, enabling DotAll in Sigil made this work: <div lang="en">(.*)</div> Do I understand correct that . now also includes new line and that's why it works?

02-04-2019, 12:19 PM	#14
DaleDe Grand Sorcerer Posts: 11,470 Karma: 13095790 Join Date: Aug 2007 Location: Grass Valley, CA Device: EB 1150, EZ Reader, Literati, iPad 2 & Air 2, iPhone 7	Sounds like a good addition to our wiki. Dale

02-04-2019, 12:53 PM	#15
patrik Guru Posts: 682 Karma: 4568205 Join Date: Jan 2010 Location: Sweden Device: Kobo Forma	Tex, this forum has some really exellent contributers, and you sure are one of them. Thank you! Will send you a pm.