How to make regex to replace 2 spaces between words, with one space?

crankypants · 10-29-2015, 07:28 AM

Sigil 0.8.7 on Windows 8.

When I paste a text file into Sigil it does lots of formatting for me, that's great. But many times I end up with 2 spaces between words, and in the Sigil preview window, these 2 spaces are not compressed into 1 space, which I thought XHTML would do, just like HTML.

So, I want to replace 2 spaces (or more) that are between words, with one single space. Example where _ is a space.

Code:

Good__morning_today.

Should end up as:

Code:

Good_morning_today.

This should not be affected.

Code:

_____<p>This is the beginning of a paragraph.</p>

DiapDealer · 10-29-2015, 08:41 AM

First off: Sigil Preview should render multiple space characters as a single space. That's the way all (x)html works. If it's not, it means that 1) the double spaces are inside a <pre> tag which indicates all whitespace is to be preserved; 2) same as #1, but "pre" is assigned through css (adobe products are notorious for this); or 3) the space characters are special no-breaking unicode characters. A fourth scenario is that spaces are being converted to   entities when pasting with formatting. Look in code view to check.

But all that aside ... the Captain Overkill in me, would use something like:

Code:

(*UCP)\b[^\S\p{Zl}\p{Zp}\n\r\t]{2,}\b

and replace what it matches with a single space character.

But that's just me.

And even that's still not going to work for situations like:

Code:

Sometimes_punctuation_like_this,__will_screw_things_up.

In that case, I'd use something like:

Code:

(*UCP)(\b|\p{P})[^\S\p{Zl}\p{Zp}\n\r\t]{2,}(\b|\p{P})

After I did the initial find and replace (no replace expression given for that one, by the way. It complicates things).

Though that may not always achieve the desired result--depending on the text. The bottom line is: don't blindly do a replace all. Step through each instance and verify the replace.

It basically looks for word boundaries (\b - made unicode aware by (*UCP)) and looks for two or more consecutive whitespace characters (not including any newlines, returns, tabs, or unicode paragraph/line separators) between them.

** And yes ... that's "NOT not whitespace" logic in there.

People who think they don't have to worry about any possible unicode characters or punctuation issues could probably get away with:

Code:

\b[^\S\n\r\t]{2,}\b

** None of my regex will work if the whitespace is being achieved with html entities.

Turtle91 · 10-29-2015, 11:29 AM

Then there is the very simplified version:

Search: " " (two spaces)
Replace: " " (one space)

When you save, it'll put all the organizing white space back in there so your paragraphs all line up.

DiapDealer · 10-29-2015, 11:43 AM

But that doesn't meet the OP's requirement that the solution must leave indentation alone. Yes, pretty-print will put indentation back when you save, but what if the user has unchecked the Clean on Save/Open option(s) or disabled pretty-print altogether? (Did I mention Captain Overkill?

)

theducks · 10-29-2015, 11:51 AM

right-click Reformat HTML: clean source

will put back any missing code indention
The space before opening Block tags (<p <div <h# ...) is for Humans, just like the blank lines between

Code:

<p>stuff</p>p>more stuff</p>

displays no different if on 2 lines, with or without indents

10-29-2015, 08:41 AM	#2
DiapDealer Grand Sorcerer Posts: 28,361 Karma: 203720150 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	First off: Sigil Preview should render multiple space characters as a single space. That's the way all (x)html works. If it's not, it means that 1) the double spaces are inside a <pre> tag which indicates all whitespace is to be preserved; 2) same as #1, but "pre" is assigned through css (adobe products are notorious for this); or 3) the space characters are special no-breaking unicode characters. A fourth scenario is that spaces are being converted to   entities when pasting with formatting. Look in code view to check. But all that aside ... the Captain Overkill in me, would use something like: Code: (UCP)\b[^\S\p{Zl}\p{Zp}\n\r\t]{2,}\b and replace what it matches with a single space character. But that's just me. And even that's still not going to work for situations like: Code: Sometimes_punctuation_like_this,__will_screw_things_up. In that case, I'd use something like: Code: (UCP)(\b\|\p{P})[^\S\p{Zl}\p{Zp}\n\r\t]{2,}(\b\|\p{P}) After I did the initial find and replace (no replace expression given for that one, by the way. It complicates things). Though that may not always achieve the desired result--depending on the text. The bottom line is: don't blindly do a replace all. Step through each instance and verify the replace. It basically looks for word boundaries (\b - made unicode aware by (UCP)) and looks for two or more consecutive whitespace characters (not including any newlines, returns, tabs, or unicode paragraph/line separators) between them. * And yes ... that's "NOT not whitespace" logic in there. People who think they don't have to worry about any possible unicode characters or punctuation issues could probably get away with: Code: \b[^\S\n\r\t]{2,}\b ** None of my regex will work if the whitespace is being achieved with html entities. Last edited by DiapDealer; 10-29-2015 at 09:11 AM.

10-29-2015, 11:43 AM	#4
DiapDealer Grand Sorcerer Posts: 28,361 Karma: 203720150 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	But that doesn't meet the OP's requirement that the solution must leave indentation alone. Yes, pretty-print will put indentation back when you save, but what if the user has unchecked the Clean on Save/Open option(s) or disabled pretty-print altogether? (Did I mention Captain Overkill? ) Last edited by DiapDealer; 10-29-2015 at 11:46 AM.

10-29-2015, 11:51 AM	#5
theducks Well trained by Cats Posts: 30,912 Karma: 60358908 Join Date: Aug 2009 Location: The Central Coast of California Device: Kobo Libra2,Kobo Aura2v1, K4NT(Fixed: New Bat.), Galaxy Tab A	right-click Reformat HTML: clean source will put back any missing code indention The space before opening Block tags (<p <div <h# ...) is for Humans, just like the blank lines between Code: <p>stuff</p>p>more stuff</p> displays no different if on 2 lines, with or without indents

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Missing spaces between words	giwqnbha	Calibre	2	10-18-2015 05:24 AM
regex - issue with spaces?	cybmole	Editor	43	12-31-2013 12:49 PM
Regex Find and Replace - Spaces	essayhead	Sigil	2	08-10-2012 07:41 PM
Troubleshooting can't make any spaces between words in my novel.	fantaxy	Amazon Kindle	2	08-03-2011 10:38 AM
RegEx: Removing Page Numbers that have Spaces	captainslow	Conversion	2	02-27-2011 04:14 PM

10-29-2015, 07:28 AM	#1
crankypants Hmm. Posts: 124 Karma: 2016606 Join Date: Oct 2015 Device: Android 4.2 Google Play Reader	How to make regex to replace 2 spaces between words, with one space? Sigil 0.8.7 on Windows 8. When I paste a text file into Sigil it does lots of formatting for me, that's great. But many times I end up with 2 spaces between words, and in the Sigil preview window, these 2 spaces are not compressed into 1 space, which I thought XHTML would do, just like HTML. So, I want to replace 2 spaces (or more) that are between words, with one single space. Example where _ is a space. Code: Good__morning_today. Should end up as: Code: Good_morning_today. This should not be affected. Code: _____<p>This is the beginning of a paragraph.</p>

10-29-2015, 11:29 AM	#3
Turtle91 A Hairy Wizard Posts: 3,313 Karma: 20171571 Join Date: Dec 2012 Location: Charleston, SC today Device: iPhone 15/11/X/6/iPad 1,2,Air & Air Pro/Surface Pro/Kindle PW & Fire	Then there is the very simplified version: Search: " " (two spaces) Replace: " " (one space) When you save, it'll put all the organizing white space back in there so your paragraphs all line up.

Advert

Advert