Affinity Publisher: replacing text with regex

[Pyanepsion]

Hello everyone,

I would like to replace all standard spaces between two figures (\d) with thin spaces using Regex.

Examples: 100 000, 54 000, etc.

How to proceed? Thank you for your explanations.

[walt.farrell]

1. In the Find panel, first make sure you have Regular Expression mode enabled in the format options for the Find field.
2. Then, copy a Thin Space from somewhere. For example, you can insert one in your document by using Text > Insert > Spaces and Tabs > Thin Space and then copy it. Or you can type U+2009 and then Alt+U, and then copy the character. (This step is needed because the "Special Characters" part of the Replace dialog doesn't include Thin Space as one of the options.)
3. In the Find field, type (\d) (\d) with a normal space in the middle.
4. In the Replace field type \1 then Paste the Thin Space from step 2, then type \2
5. Click the Find button, then the Replace button.

[Old Bruce]

Find

(\d) (\d)

That is a regular space in there

Replace

\1thinspace\2

You have to copy paste the thin space from somewhere into the replace field.

[Pyanepsion]

Thank you, Walt Farrell and Old Bruce.

 

[David in Яuislip]

I didn't need regex, it worked set to normal

Top text is original
Centre text I did manually - I got the thin space from the Glyph Browser
Bottom text done by Search/Replace

Two caveats:
   the reduced space width sort of confuses the layout
   be careful with Replace All, it isn't constrained to the selected Text frame but jumps to the top one and changes that too. 

[walt.farrell]

2 minutes ago, David in Яuislip said:

I didn't need regex, it worked set to normal

But @Pyanepsion did not want to replace all the spaces with thin spaces; only the ones between two digits.

True, that doesn't require regex either, but you'll be hitting Find a lot to move past the ones you don't want to change if you use a normal operation rather than regex.

[David in Яuislip]

@walt.farrellYou're quite correct, forget my drivel

[Old Bruce]

24 minutes ago, David in Яuislip said:

be careful with Replace All, it isn't constrained to the selected Text frame

Sadly it is not constrained by the text flow either. It is set to search and replace in ALL of the text in the document.

[Pyanepsion]

9 hours ago, David in Яuislip said:

I didn't need regex, it worked set to normal

Hello, David. That’s right. Regex (regular expression) can often be avoided. I would advise you, however, against using spaces other than what they were designed for non-breaking spaces, breakable spaces, thin spaces, etc. to avoid problems at the printing stage.

Digging into the question of regexes, I find an anomaly. The class \w (alphanumeric characters) is well recognized, as well as \d (numeric characters), but not \a (alphabetic characters only). Fortunately, [a-z] can be used instead of \d, but this is a surprising omission.

[lacerto]

(...)

[Pyanepsion]

The absence of \a cannot be deliberate, as this would be a major design error.

In our business, at least most of us who use DTP software, it is very common to have to rectify cut and paste errors that have added a non-breaking space.

Example: In orange, a lot of spaces found by \w that should not be changed. Only the blue spaces, usually found by \a, must be corrected.

 

[walt.farrell]

10 hours ago, Pyanepsion said:

The absence of \a cannot be deliberate, as this would be a major design error.

There are many dialects of regular expressions, with slightly different syntaxes and meanings. It is common to have to learn how to transform a regex that works in one application into one that will work in another application, with a different regex dialect.

But considering \a: In most RegEx dialects that I'm used to, the it represents the "BEL" (alarm) character. That includes in Perl and in the Boost library operating in Perl mode (which is relevant for Publisher, I think).

I don't think I've ever seen one where \a is anything other than BEL. Here's a good RegEx reference, by the way, to referring to characters: http://www.regular-expressions.info/refcharacters.html

It also has a good reference for characer classes. The closest to what you want, I think, is [:alpha:] which I think/ can work in Boost (Publisher) in some cases.
http://www.regular-expressions.info/refcharclass.html

[Pyanepsion]

\a is indeed originally found more in Vim, a Linux text editor oriented in the manipulation of text files from which most other text editors are inspired.

Each variant of Regex is in fact intended for a particular use. However, we work on text. We should therefore be able to easily perform operations such as detecting inappropriate spaces in a string of characters including punctuation, numbers, symbols, superscripts, indices, etc. This is impossible at the moment.

The main programming languages, even the most text-oriented ones such as Perl or its successor Python, have unfortunately been developed by programmers who use a mother tongue that does not know all the subtleties of accentuation, and they are not literary. I am afraid that this issue can only be resolved when Serif allows third-party companies to create extensions and provides them with the necessary documentation.

Quote

Elle était : le Petit Chaperon rouge. Ilª ne le savait pas.

10 spaces must be standard, 1 is a non-breaking space (in front of the colon) and a non-breaking space is missing before the note call.
[[:alpha:]] [[:alpha:]] : 7 occurences (83 calculation steps)
[[:word:]] [[:word:]] : 7 occurences (83 steps of calculation)
(\w) (\w) : 7 occurences (166 steps of calculation)

[walt.farrell]

2 hours ago, Pyanepsion said:

We should therefore be able to easily perform operations such as detecting inappropriate spaces in a string of characters including punctuation, numbers, symbols, superscripts, indices, etc. This is impossible at the moment.

Sorry, but I don't understand why it is impossible. And my apologies, but I also do not understand how your example demonstrates this, or what your example demonstrates.

[lacerto]

(...)

[walt.farrell]

25 minutes ago, Lagarto said:

so the following would find all special spaces that Affinity apps recognize and would replace them all with a regular space,

If you want to replace all special spaces with a normal space, it's even easier. Just replace all spaces with a regular space. Search for \s and replace with a space.

[lacerto]

(...)

[Pyanepsion]

5 hours ago, walt.farrell said:

If you want to replace all special spaces with a normal space, it’s even easier. Just replace all spaces with a regular space. Search for \s and replace with a space.

The usual problem is exactly the opposite. In neat French typography, there are 14 types of spaces, which vary according to punctuation, number, certain types of words, etc. All are of course available in Affinity Publisher.

It is common to find inappropriate spaces in client texts. They are usually the result of poor copy and paste. They are easy to replace, but this currently requires a lot of attention to detail and time. Better ergonomics would save a lot of time for all users whose language uses complex typography.

There are currently 3 possibilities:

1. Checking each of the space types one by one in the whole text. Very long and with risks of careless mistakes.
2. Using a rather complicated Regex routine. Affinity Publisher only offers the minimum here. Long, but very precise.
3. Using a third-party extension, which Serif still does not allow. Very fast, very precise, and error-free.