Interpreting non-ASCII characters as integers

[niruvt]

Try this code -

\documentclass{article}
\usepackage{marathi}
\usepackage[backend=biber, style=authoryear]{biblatex}
\begin{filecontents}[overwrite]{\jobname.bib}
@article{गौरी,
  author       = {मोघे, गौरी},
  title        = {तन्त्रशास्त्राचा परिचय},
  journaltitle = {आलोक मराठी नियतकालिक},
  date         = {२०२१},
  series       = {३},
  pages        = {१-५},
}
\end{filecontents}
\addbibresource{\jobname.bib}

\begin{document}
\cite{गौरी}

\printbibliography
\end{document}

One might notice that pp. and In: are missing in the output. This is because ३ and १-५ are understood as characters and not as numerals, but well technically they are numerals. Changing ३ to 3 and १-५ to 1-5 works. This forces users to use Latin numerals in their .bib files. For Marathi at least Devanagari numerals are shown even with the Latin numerals in the code because of the conversion programs by XeLaTeX and LuaLaTeX, but I am not sure if it is true for other languages, which means in the lack of a conversion program the output will show Latin numerals. Can something be done to make biber understand that १, २, ३, ४, ५, ६, ७, ८, ९ are numerals? (At least when these come in the .bib file of Marathi/Hindi/Sanskrit/Nepali.)

[sieversMartin]

Have you tried to configure numbers with \DeclareNumChars?

[niruvt]

Nope. How to do that?

[sieversMartin]

Please consult p. 262 of the documentation. Basically you can give a list of characters to be considered as a number.

[sieversMartin]

\DeclareNumChars{१,२,३,४,५,६,७,८,९}

[niruvt]

Thanks. Closing.

[moewew]

Note that \DeclareNumChars does not take a comma-separated argument. It just takes an undelimited list of characters. So you'd probably want

\DeclareNumChars{१२३४५६७८९}

[niruvt]

@moewew Thanks, but then is it understood as one number or every character is understood as different a separate string?

[moewew]

Each character is processed and treated separately. The argument of \DeclareNumChars is treated as an undelimited list of single characters.

[niruvt]

Okay thanks.