search whith Æ ou Œ

[Doum]

hi,
When i run search with title:"~Ætern" or title:"~ætern" i find all my books with this type of character ( lettre soudée in french, but in english i don't know the translation , welded letter perhaps ) (idem with Œ)

But this character in french meens "ae" or "oe" with 2 letters, so i would like to find also all books like "Ad Vitam Aeternam" or "Oedipe" .
Have you an idea ?

The strange thing is the function strtolower PHP("rÆd") give rÆd and not ræd, but calibre (pyton?) do the good job and find lower/upper title. (like the sql LIKE ).

[Quoth]

I don't quite understand where or how you are searching as that makes a difference. Calibre main screen GUI, command line, a program script etc.

Titles in filenames probably won't use Æ etc.


Some background for searching in an eBook.
Æ Œ and similar combinations can either be a ligature (any two letters run together such as fi sometimes) or a single letter, a diagraph, made from two letters. A diagraph is entered as single character in the source text.

In the modern French alphabet, æ (called e-dans-l'a, 'e in the a') is used to spell Latin and Greek borrowings like curriculum vitæ, et cætera, ex æquo, tænia, and the first name Lætitia. It is mentioned in the name of Serge Gainsbourg's song Elaeudanla Téïtéïa, a reading of the French spelling of the name Lætitia: "L, E dans l'A, T, I, T, I, A."

However ebooks are usually coded in HTML. Most ebooks manage Æ æ Œ œ as characters fine, but the publisher designer may use an HTML entity such as Æ æ Œ or œ

So you may have to search for the actual character or an HTML entity: https://en.wikipedia.org/wiki/List_o...ity_references

[Doum]

look google book on https://www.google.com/search?q=r%C3...=gws-wiz-books

[Quoth]

I know what they are, I just don't understand exactly where and how you are searching.

æ in a title can be searched for on Calibre GUI.

Also I can easily type them as I use Linux. Compose Key a e gives æ etc. Common in old Anglo-Saxon names.

[kovidgoyal]

calibre's search will match accented to unaccented characters and ignore case as well controllable via Preferences->Searching. However it doesnt do the transform you are asking for. Searching isnt language specific and there is no general rule that says ae and æ should be the same.

[Sarmat89]

Originally Posted by kovidgoyal

Searching isnt language specific and there is no general rule that says ae and æ should be the same.

Calibre already does that for normal search, but not for regex one.

[Quoth]

In general ae is two characters and æ is one. The a e can be in words never spelled with an æ such as Fae, so ae and æ shouldn't be treated the same, unlike say i and ï (dais) or e and é (cafe) etc which the exact same word can be spelled without the accents. Admittedly some words in English can be spelled with a diaeresis or not, such as mediaeval.

[Sarmat89]

Originally Posted by Quoth

In general ae is two characters and æ is one.

Not unless you have a tailored collation for some Scandinavian language. By default, it is A+E.

[Quoth]

Originally Posted by Sarmat89

Not unless you have a tailored collation for some Scandinavian language. By default, it is A+E.

I was referring to spelling and typography were the source can be separate a and e or there can be a æ single character. What any particular search or sort system/program does is a separate issue.

Checking with caja file manager sort: aertel is always before ærtel, no matter the creation order. Both are after adept and before afro.

My GUI Language setting is English UK, Region & Time is English Ireland. (Spelling checkers have no English Ireland setting). Installed Languages are English UK, English Ireland and English USA.

In general the sort order will be different on some 8 bit ASCII DOS systems with æ in the upper 127 characters.

[Doum]

The use case is not to do a search with æ or œ which always succeeds (if exact character exists) but to do a search with 2 letters : ae or oe and also find the one letter æ or œ .

I readily recognize that this request is superfluous because I only need to make 2 requests or instead of one. It's purely theoretical
Anyway thank you for your comments

For the fun (with my keyboard) :
œ => Alt. + 0156
Π=> Alt. + 0140
æ => Alt. + 145 ou Alt. + 0230
Æ => Alt. + 146 ou Alt. + 0198