\IfDecimal returning Number to big

Question

\IfDecimal returns an error number to big with large decimal. My goal is to print a percent rounded to to places if it is a number, otherwise just print the value of the parameter. Is there away around this size limit or is there another way around this given that I don't yet know whether I have a parsible number? The \SomePerc is a simplified macro that I am trying to use. Input to SomePerc is either a floating point or a "-" (as dash not negative)

\documentclass[]{article}
\usepackage{siunitx}
\usepackage{xstring}
\newcommand\SomePerc[1]{
    \IfDecimal{#1}
    {\qty[  exponent-mode = fixed,
        fixed-exponent = 0,
        drop-exponent=true, 
        drop-zero-decimal=true,
        round-mode=places,
        round-precision =2]{#1e2}{\percent}
    }
    {#1}
}
\begin{document}
\IfDecimal{0.05479140921}{true}{false}
\IfDecimal{0.0547914092}{true}{false}
\num{0.054791409210}
\SomePerc{0.05479140921}
\SomePerc{0.0547914092}
\end{document}

The latex compiler says:

Number too big. \IfDecimal{0.05479140921}
Number too big. \SomePerc{0.05479140921}

Probably some xstring bug. I don't think this package is very actively developed though, so either roll your own or use something else e.g.. egreg's answer based on l3regex https://tex.stackexchange.com/questions/463814/check-whether-string-represents-valid-number-for-pgf# — user202729, May 20 '23 at 00:32
what exactly is the kind of input your \SomePerc is supposedly accepting? — user691586, May 20 '23 at 09:02

score 2 · Accepted Answer · answered May 20 '23 at 17:57

The following defines \ifdecimal, an expandable, fastish test that works for arbitrarily big numbers and places. It works for both , and . as the decimal separator, and arbitrary many signs. Everything that has at least one digit, at most one decimal separator, and no spaces is considered a decimal number. The input is fully expanded then stringified, the result of which has to match the aforementioned rules.

\documentclass{article}
\makeatletter
% needed to get a space inside of a definition
\newcommand\exparg@o[2]{\expanded{\unexpanded{#1}\expandafter}\expandafter{#2}}
% the front facing macro
\newcommand\ifdecimal[1]
  {%
    \expandafter\ifdecimal@space\detokenize\expanded{{#1}}\STOP
    \ifdecimal@true
  }
% true and false branching
\let\ifdecimal@true@firstoftwo
\long\def\ifdecimal@false#1\ifdecimal@true#2#3{#3}
% check has no spaces
\exparg@o{\def\ifdecimal@space#1\STOP}%
  {%
    @firstofone{\ifdecimal@space@#1\STOP\ifdecimal@false} \STOP % <- space
    \ifdecimal@sign#1\MARK,\MARK
  }
\def\ifdecimal@space@#1 #2\STOP{}
% strip leading signs
\def\ifdecimal@sign#1%
  {%
    \ifdecimal@sign@#1-\ifdecimal@sign@true+#1\ifdecimal@sign@true
      +-\ifdecimal@dot#1%
  }
\def\ifdecimal@sign@#1+-{}
\def\ifdecimal@sign@true#1+-\ifdecimal@dot#2{\ifdecimal@sign}
% turn one , into a . and then remove one decimal marker
\def\ifdecimal@dot#1,{\ifdecimal@dot@#1.}
\def\ifdecimal@dot@#1.{\ifdecimal@healmark\ifdecimal@empty#1}
\def\ifdecimal@healmark#1\MARK#2\MARK{#1\STOP}
% check is non-empty
\def\ifdecimal@empty#1\STOP
  {%
    \ifdecimal@empty@\MARK#1\STOP\ifdecimal@false\MARK\STOP
    % the {0 \ifdecimal@done} works as the right delimiter to the input argument
    \ifdecimal@digits#1{0 \ifdecimal@done}%
  }
\def\ifdecimal@empty@#1\MARK\STOP{}
% check only digits left
\def\ifdecimal@digits#1%
  {%
    \ifnum9<1#1
    \else
      \expandafter\ifdecimal@false
    \fi
    \ifdecimal@digits
  }
\def\ifdecimal@done#1\ifdecimal@digits{\fi}
\let\MARK\relax
\let\STOP\relax
\makeatother
% simple tests whether the macro works
\newcommand\ASSERT[2]
  {%
    \begingroup
      \edef\tmpa{#2}%
      \def\tmpb{#1}%
      \expandafter
    \endgroup
    \ifx\tmpa\tmpb
    \else
      \GenericError{}{Assertion failed for \detokenize{#2}}{}{}%
    \fi
  }
\ASSERT{true}{\ifdecimal{0.}{true}{false}}
\ASSERT{true}{\ifdecimal{.0}{true}{false}}
\ASSERT{true}{\ifdecimal{0.0}{true}{false}}
\ASSERT{true}{\ifdecimal{00}{true}{false}}
\ASSERT{true}{\ifdecimal{+-+-+-+-00}{true}{false}}
\ASSERT{false}{\ifdecimal{.}{true}{false}}
\ASSERT{false}{\ifdecimal{+}{true}{false}}
\ASSERT{false}{\ifdecimal{-}{true}{false}}
\ASSERT{false}{\ifdecimal{+-}{true}{false}}
\ASSERT{false}{\ifdecimal{+-.}{true}{false}}
\ASSERT{false}{\ifdecimal{0+}{true}{false}}
\ASSERT{false}{\ifdecimal{0.0.0}{true}{false}}
\ASSERT{false}{\ifdecimal{a}{true}{false}}
\newcommand\five{5}
\ASSERT{true}{\ifdecimal{\five}{true}{false}}
\ASSERT{true}{\ifdecimal{\five .\five}{true}{false}}
\ASSERT{true}{\ifdecimal{\empty .\five}{true}{false}}
\ASSERT{false}{\ifdecimal{\five .\relax}{true}{false}}
\stop

\IfDecimal returning Number to big

1 Answers1