check for numeric content in a shell script (FreeBSD sh)

[parv at pair.com]

in message <20100624192256.GF557 at libertas.local.camdensoftware.com>,
wrote Chip Camden thusly...
>
> On Jun 24 08:39, Parv wrote:
> > in message <20100624183407.GA49923 at holstein.holy.cow>,
> > wrote parv at pair.com thusly...
> > >
> > > #  Matches a number, either positive (without '+' sign) or
> > > #  negative, which is either a whole number; or a real number
> > > #  ending with decimal point, or a real number with or without
> > > #  leading digits before the decimal point.
> > .                 ^
> > .                 ^  plural
> > > ^
> > > -?
> > > (
> > >   [0-9]  [.]? [0-9]*
> > >     |
> > >   [0-9]? [.]  [0-9]+
> > .        ^
> > .        ^  oops
> >
> > Please change the immediately above regex portion to ...
> >
> >   [0-9]* [.]  [0-9]+
...
> We still need to be able to handle numbers without a decimal.

First alternative above handles that ...

  [0-9]   #  Match 1 digit,
  [.]?    #  followed by an optional decimal,
  [0-9]*  #  followed by any number of optional digits.


> Try this:
>
>       [0-9]*\.?[0-9]+

If it is really /^[0-9]*\.?[0-9]+$/, then it does not match
a negative number or a number ending with a decimal (e.g. 8.).


> The question mark says "0 or 1"
> >
> > > )
> > > $

Annotated regex now is ...

  ^       #  Anchor at the beginning of string;
  -?      #  followed by an optional -ve sign;
  (       #  start grouping|alternatives;
    [0-9]   #  match 1 digit,
    [.]?    #  followed by an optional decimal,
    [0-9]*  #  followed by any number of optional digits;
     |      #  OR,
    [0-9]*  #  match any number of optional digits,
    [.]     #  followed by 1 decimal point,
    [0-9]+  #  followed by 1 or more digits;
  )       #  end of grouping;
  $       #  anchor at the end of the string.


  - parv

--