small roman numerals in header – regex help

[Holden100]

I am trying to search and replace while converting PDF to Epub but stuck at how to take care of small Roman numbers (from i to xxxi) that appear as headers in the prologue – the even numbers are: xii<br>

prologue<br> etc while the odd numbers are prologue<br>xix<br> etc.

Clearly [xvi] does not help

[theducks]

You are looking for a sequence of characters contained in the SET [xvi]

Find: [xvi]{1,3}

[Holden100]

Originally Posted by theducks

You are looking for a sequence of characters contained in the SET [xvi]

Find: [xvi]{1,3}

Searching for this string [xvi]{1,3} includes all those words where i, x or v also appear.

Actually, it's a book and the small roman numerals are used only for the prologue pages – which run from 1 to 31, or i to xxxi

The string [xvi]{1,3} ends up finding over 34,000 occurrences. I am only looking to find around 31.

I don't know regex at all – but clearly it would need something like /b or something to narrow the searches down to only such cases where the words containing them are at the least 1 digit (from i, v, x) through two (ii, iv, vi, ix, xi, xx), three digits (iii, vii, xii, xiv, xvi), four digits (viii, xiii, xvii, xxii, xxiv, xxvi, xxix) and five digits (xviii, xxiii, xxvii) . or at the most six digits (xxviii) long

[Holden100]

So basically I am looking to find words composed either, all or some repetition of letters [xvi] that could be from one to six letters long!

[theducks]

Originally Posted by Holden100

So basically I am looking to find words composed either, all or some repetition of letters [xvi] that could be from one to six letters long!

I changed the {1,6} to {1,3}, due to faulty thinking
The search I actually used, included the tags that existed in my document..
Again, I faild by overly simplifing

BTW I only make changes S&R using the editor. I don't try and do this kind of stuff during conversion

[Holden100]

But clearly, something else is also required as simply doing [xvi]{1,6} still shows up over 34,000 results, when it should only be finding about 31 or so.

In fact, the search is locating almost ANY word with the letter x, v or i whereas we need to be looking ONLY for letters composed of these letters and no other letter.

[lumpynose]

Originally Posted by theducks

BTW I only make changes S&R using the editor. I don't try and do this kind of stuff during conversion

Converting from PDF is often a minefield with lots of things going bad. So this suggestion is something to pay attention to. Try doing a conversion with nothing else going on (no replacements), then open the epub in the editor and see how much damage was done. Other epub editors might also be helpful; I use sigil but I don't know the pros and cons of sigil vs calibre's editor.

[Holden100]

But surely there must be some way of writing a regular expression for small roman numerals from i to xxxi?

[Holden100]

Sorry, not clear. Please help.

I am trying to search and replace while converting PDF to Epub but stuck at how to take care of small Roman numbers (from i to xxxi) that appear as headers in the prologue – the even numbers are: "ii<br>prologue<br>" etc while the odd numbers are prologue"<br>xix<br>" etc.

Clearly [xvi]+ does not help as it ends up highlighting all words which have any of these letters.

Actually, it's a book and the small roman numerals are used only for the prologue pages – which run from 1 to 31, or i to xxxi

The string [xvi] ends up finding over 34,000 occurrences. I am only looking to find around 31.

I don't know regex at all – but clearly it would need something like /b or something to narrow the searches down to only such cases where the words containing them are at the least 1 digit (from i, v, x) through two (ii, iv, vi, ix, xi, xx), three digits (iii, vii, xii, xiv, xvi), four digits (viii, xiii, xvii, xxii, xxiv, xxvi, xxix) and five digits (xviii, xxiii, xxvii) . or at the most six digits (xxviii) long

[BetterRed]

Moderator Notice

Please don't post the same issue in multiple threads - See #3 - Duplicates in MR Guidelines

I moved your post in Blurr's thread into here

BR

[davidfor]

Try:

\b[ivx]+\b

The "\b" is for the word delimiter. The only issue I can think of is if there is one of those letters somewhere else in the book. And the following should fix if I read what you are saying correctly:

\b[ivx]+<br

[Holden100]

Originally Posted by davidfor

Try:

\b[ivx]+\b

The "\b" is for the word delimiter. The only issue I can think of is if there is one of those letters somewhere else in the book. And the following should fix if I read what you are saying correctly:

\b[ivx]+<br

Thank you so much! I think this works!

[Holden100]

Originally Posted by BetterRed

Please don't post the same issue in multiple threads - See #3 - Duplicates in MR Guidelines

I moved your post in Blurr's thread into here

BR

Apologies