MathGroup Archive 2011

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Apply a rule to an expression only once

  • To: mathgroup at smc.vnet.net
  • Subject: [mg117080] Re: Apply a rule to an expression only once
  • From: Guido Walter Pettinari <coccoinomane at gmail.com>
  • Date: Wed, 9 Mar 2011 06:59:39 -0500 (EST)
  • References: <ij2v15$7vn$1@smc.vnet.net> <ij5mp0$ji5$1@smc.vnet.net>

Thank you very much for all the answer!  All of them were very
interesting and taught me something.   In the end I adopted Ray's
solution, with a slight modification in order to prevent an error if
the rule doesn't apply:

If [
 Position[expr, rule[[1]]] === {},
 expr,
 MapAt[ReplaceAll[#, rule] &, expr, Position[expr, rule[[1]]][[1]]]]
]

Cheers,

Guido



On Feb 12, 10:18 am, Ray Koopman <koop... at sfu.ca> wrote:
> On Feb 11, 1:20 am, Guido Walter Pettinari <coccoinom... at gmail.com>
> wrote:
>
>
>
>
>
>
>
>
>
> > Dear Andrzej,
>
> > thank you for your answer.  Your example works indeed, but I would
> > rather prefer not to modify my original rule, which is much more
> > complex than the one I quoted in the example.  I want the rule to
> > stay the most general as possible, since it is used elsewhere in my
> > package and a redefinition will break it.
>
> > My package transforms mathematical equations from a real [x,y,z]
> > space to a Fourier [kx,ky,kz] space.  The main ingredient is a rule
> > that applies the necessary derivative & coordinate transformations on
> > single terms (e.g. Derivative[1,0,0][ f [x,y,z] ] -> I kx f [kx,ky,kx] =
).
> > This works perfectly fine on linear terms, but things get more
> > involved when dealing with quadratic terms.   In fact, the Fourier
> > transform of a quadratic term, say f [x,y,z] g [x,y,z], is a
> > convolution term of the form C [ f [k1,k2,k3] g[q1,q2,q3] ],
> > where C[] denotes a convolution integral over k1,k2,k3 and q1,q2,q3.
>
> > Aside from mathematical details, I would like to know a way to
> > change the expression:
>
> > expr = f [ x ] g [ x ] h [ x ] .....
>
> > into
>
> > f [ k1 ] g [ k2 ] h [ k3 ] ....
>
> > by applying in a proper way the following rule:
>
> > rule[k_] := f_@x :> f@k .
>
> > No redefinitions of the rule are allowed :-)
>
> > I thought that a solution could be an option for Replace (say,
> > OnlyFirst) that applies the rule only to the first occurrence of
> > the pattern:
>
> > expr // Replace [ #, rule[k1], OnlyFirst->True ]& //
> >         Replace [ #, rule[k2], OnlyFirst->True ]& //
> >         Replace [ #, rule[k3], OnlyFirst->True ]& // ...
>
> > but I do not know how to implement the "OnlyFirst" bit.
>
> > Please note that (i) I do not care which function gets which
> > dependence, e.g. f[k2]g[k3]h[k1] is equally good for me, and
> > (ii) I do not know the names of the functions, I only recognize
> > them by their coordinate dependence (usually x,y,z).
>
> replaceFirst[expr_, rule_] := MapAt[ Replace[#, rule]&,
>              expr, Position[expr,rule[[1]]][[1]]]
>
> expr = f[x] g[x] h[x];
>
> rule[k_] = f_[Except[k1|k2|k3]] :> f[k];
>
> Fold[replaceFirst, expr, rule/@{k1,k2,k3}]
>
> f[k1] g[k2] h[k3]



  • Prev by Date: Re: evaluation-- one or many levels, your thoughts?
  • Next by Date: Re: evaluation-- one or many levels, your thoughts?
  • Previous by thread: Re: Algebraic substitution with PolynomialReduce
  • Next by thread: Re: memory leak in Mathematica application using NET/Link