On Thu, 18 Jun 2009, David Mertens wrote:
> Most everybody on this list probably understands the power of regexp's for
> string processing. Has anybody ever given thought to creating a vector
> regexp (I would call it 'subsetting') for codifying complex slices?
I'm not sure how one would use this in practice but...
>
> Here's what I've come up with so far:
>
> Basic Metacharacters
> ====================
> ^ Match the beginning of a vector
> $ Match the end of a vector
> | Alternation
> () Grouping
> . Match a single number
>
> Numerical Ranges
> ================
> (a,b) match any number c such that a < c < b
> (a,b] match any number c such that a < c <= b
> [a,b) match any number c such that a <= c < b
> [a,b] match any number c such that a <= c <= b
>
> Quantifiers
> ===========
> * Match 0 or more times
> + Match 1 or more times
> ? Match 1 or 0 times
> {n} Match exactly n times
> {n,} Match at least n times
> {n,m} Match at least n but not more than m times
>
> Slicing
> =======
> s() Include this group in the resulting slice.
>
> Metacharacters
> ==============
> p matches a positive number
> P matches a positive number or zero
I'd swap that
P matches a negative number or zero
> n matches a negative number
N matches a positive number or zero
so that n is not N, p is not P, as it is in regexps
> N matches a negative number or zero
> o matches zero
> O matches nonzero
Like this is: o is not O
> M matches a local maximum
> N matches any character that is not a local maximum
conflicts with negative.
> m mathces a local minimum
> n matches any character that is not a local minimum
> M{n} matches a local maximum near index n
> m{n} mathces a local minimum near index n
> \M matches the global maximum
> \m matches the global minimum
>
> Zero-width Assertions
> =====================
> \c Match a zero crossing
> \c+ Match a positively sloped zero crossing
> \c- Match a negatively sloped zero crossing
> \c=a Match when the time-series crosses the value a
> \c=a+ Match when the time-series crosses the value a with positive slope
> \c=a+ Match when the time-series crosses the value a with negative slope
> \dN Match where the Nth numerical derivative crosses zero (also for +- and
> =)
I'd add
S(...) sum the elements matched
t(...) times: product of the elements matched
b(...) Bounds of the elements matched
l(...) length of match, i.e a count
i is integer
I is noninteger
c is complex
Possibly
f{ } f($x) is true in the Perl sense (nonzero), where $x works like
$a and $b in the sort function, and perldl code goes in {...}
And if you have operators to move forward and backwards you've got the
best part of a Turing machine!
I don't know how one might apply Is a positive maximum, is a negative maximum.
>
Hugh
_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
