RegEx Helper as a plugin can offer some help (assuming that RegEx Helper is already installed):

Open your file in Notepad++
Set the language to Php
Launch the plugin using Ctrl+F12
Type in the “Regular Expression” window: ({{)|(}})
Turn on the highlighting using ‘Mark’ button.

Tested using this Moustache template:
<strong>Hello, {{firstname}}</strong><br>
<p>You are visitor number {{visitorNumber}}.</p>

Hello, @scott-sumner and All,

I’ve studied the general case of searching a specific character, ONLY IF, located inside a range of characters with delimiters.

Now, two cases are possible :

Case A : an area with a same starting and ending character, as, for instance, '.....' or "....."

Case B : an area with a different starting and ending character, as, for instance, (.....), [.....], {.....} or <.....>

Notes :

For our discussion, we are supposed to look for the colon character

For case A, I chose the double quotes delimiter ", as common boundary

For case B, I chose the start delimiter < and the end delimiter >

Let’s begin with the easier form !

Case B : A possible regex would be :

SEARCH :(?=[^<\r\n]*>)

REPLACE Any string or character, even EMPTY

Note that this regex looks for a colon character, ONLY IF followed by a range, possibly empty, of characters, different from the first delimiter < and from the EOL characters \r and \n , till the ending delimiter >

On the test example, below, the regex finds all colon characters, located inside the <.....> areas exclusively ( the ones which are underlined ) Fine !

1:23:<This:is a: tiny>text:to :see<if :my :logic:>is: correct:<I: hope:that>: all: will:be<::fine,: indeed>: ! :<:>78:9 ¯ ¯ ¯ ¯ ¯ ¯ ¯ ¯¯ ¯ ¯

Now, in case A, the annoying thing is that it’s impossible to distinguish the two delimiters ! So, we’re going to cheat a bit ! First, we’ll replace, for instance, any area "....." by an oriented area as, for instance, #"....."@. Of course, these new boundaries must be absent from the present contents of the file !

So, assuming the original text :

1:23:"This:is a: small"text:to :see"if :my :logic:"is: correct:"I: hope:that": all: will:be"::fine,: indeed": ! :":"78:9

The simple S/R :

SEARCH ".*?"

REPLACE #$0@

would get the following text :

1:23:#"This:is a: small"@text:to :see#"if :my :logic:"@is: correct:#"I: hope:that"@: all: will:be#"::fine,: indeed"@: ! :#":"@78:9

Accordingly, the correct regex becomes :

SEARCH :(?=[^#\r\n]*@)

REPLACE Any string or character, even EMPTY

Again, only the colons, inside the areas, which are underlined, are matched by the regex !

1:23:#"This:is a: small"@text:to :see#"if :my :logic:"@is: correct:#"I: hope:that"@: all: will:be#"::fine,: indeed"@: ! :#":"@78:9 ¯ ¯ ¯ ¯ ¯ ¯ ¯ ¯¯ ¯ ¯

To end with, use the simple regex :

SEARCH #|@

REPLACE Empty

in order to get the original areas "....."

Et voilà !

Cheers,

guy038

@guy038

Thanks…what confused me was that it sounded like you were saying that tabstops were always 4 columns…your most-recent post clears that up.

@Vini-Dope

Well, if you can’t share it globally you probably shouldn’t share it with me.

I made a much bigger dataset (100000 lines of random text) and was able to duplicate your bad result (all text selected after the Mark operation). To be honest I don’t know what is going on with that…maybe it is just “too much” for the regex engine…

Here’s something else you can try. Take your word list and replace all of the line endings between the words with the vertical bar: |

You can do this by searching the word list for \R and replacing with |– regular expression search of course.

Thus you’ll end up with something like this: Xxxxxxxx|Zzzzzzzz|Ccccccc|Vvvvvvv|Bbbbbb|Nnnnnn

Take that long string and copy and paste it into the Find-what zone and do a Mark (or a Find) as described in Steps 2 and 3 above. This should find your words in your large document.

Assuming your words-to-find average 10 characters in length, you can in theory use this technique on roughly 200 words (the limit of the Find-what zone is 2046 characters. I did not try that many with my experiment, but in theory at least…