MathGroup Archive 2010

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

Search the Archive

Re: How to short-circuit match failure?

  • To: mathgroup at smc.vnet.net
  • Subject: [mg114258] Re: How to short-circuit match failure?
  • From: Andrea <btlgs2000 at gmail.com>
  • Date: Mon, 29 Nov 2010 06:06:55 -0500 (EST)
  • References: <ictfq8$mck$1@smc.vnet.net>

On Nov 28, 12:52 pm, kj <no.em... at please.post> wrote:
> When defining a function as a sequence of rules (with SetDelayed),
> I often want to have messages emitted if the arguments do not have
> the proper form, and then *stop* trying any remaining rules, but
> I don't know how to do the latter.
>
> Here's a silly example:
>
> ClearAll[foo]
> foo::toolong = "List is too long";
> foo::nolist = "First argument is not a list";
> foo::nargs = "foo called with `1` argument(s); 2 expected";
> foo[x_List /; Length[x] < 3, y_] := {#, y} & /@ x
> foo[x_List, y_] /; Message[foo::toolong] = Null
> foo[x_, y_] /; Message[foo::nolist] = Null
> foo[x___] /; Message[foo::nargs, Length[{x}]] = Null
>
> The function foo takes as its first argument a list with length no
> greater than 2.  But see what happens when foo[{1, 2, 3}, 3] is
> evaluated:
>
> In[86]:= foo[{1, 2, 3}, 3]
> During evaluation of In[86]:= foo::toolong: List is too long
> During evaluation of In[86]:= foo::nolist: First argument is not a list
> During evaluation of In[86]:= foo::nargs: foo called with 2 argument(s)=
; 2 expected
> Out[86]= foo[{1, 2, 3}, 3]
>
> The result in Out[86] is as desired, but spurious messages were
> emitted.  It's easy to see why.  During the evaluation of foo[{1,
> 2, 3}, 3], *all* the rules associated with foo are tried, even
> though the third and fourth ones should not be.
>
> I can prevent this from happening by letting the second rule's
> match succeed; e.g. by defining it like this:
>
> foo[x_List, y_] := (Message[foo::toolong]; Null)
>
> ...but now evaluating foo[{1, 2, 3}, 3] no longer produces the right
> final result (it produces Null, instead of foo[{1, 2, 3}, 3]).
>
> How can I define the second rule for foo so that the third and
> fourth ones are not tried, while the final value for foo[{1, 2,
> 3}, 3] remains as foo[{1, 2, 3}, 3]?
>
> TIA!
>
> ~kj

I kj,it is an interesting question to me because I never defined
functions with messages.
It seems to me that there is no way to skip other definitions once you
match the right error.
So I would resolve the problem in the following way:
make the right definition e.g

foo[x_List /; Length[x] < 3, y_] := {#, y} & /@ x;

followed by definitions each matching with one particular error:

foo[x_,y_]/;(condition_to_match &&
(Message[one_message];False) ):=Null


In your situation

foo[x_List, y_] /; (Message[foo::toolong]; False) = Null;

foo[Except[_List], y_] /; ( Message[foo::nolist]; False) = Null;

foo[x___] /; (Length[List[x]] != 2 && (Message[foo::nargs,
Length[{x}]]; False)) = Null;

It seems to me it works but perhaps there is a more elegant way
Andrea


  • Prev by Date: Re-virginating Manipulates?
  • Next by Date: Re: How to short-circuit match failure?
  • Previous by thread: How to short-circuit match failure?
  • Next by thread: Re: How to short-circuit match failure?