MathGroup Archive 1998

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

Search the Archive

Re: Inconsistencies in pattern matching.

  • To: mathgroup at smc.vnet.net
  • Subject: [mg12992] Re: [mg12937] Inconsistencies in pattern matching.
  • From: "Jrgen Tischer" <jtischer at col2.telecom.com.co>
  • Date: Sun, 28 Jun 1998 02:52:14 -0400
  • Sender: owner-wri-mathgroup at wolfram.com

Ok Sean,
let's try to convince you of the unthinkable. First the "definition" of
ReplaceAll (Book, online help on ReplaceAll):

ReplaceAll looks at each part of expr, tries all the rules on it, and
then
goes on to the next part of expr. The first rule that applies to a
particular part is used; no further rules are tried on that part, or on
any
of its subparts.

What it says is that ReplaceAll is traversing the tree structure of the
expression breadth first. What you don't like so much is that
Mathematica starts at the root. But this IS very consistent. How should
Mathematica know
that in the case of {x1,y1}/.{x_,y_}->{x,Log[y]} you want it start at
the root and in the case of {{x1,y1},{x2,y2}}/.{x_,y_}->{x,Log[y]} not?
By the
way this is what restrictions as you used one are good for.

Satisfied?

Jrgen

-----Original Message-----
From: Sean Ross <seanross at worldnet.att.net> To: mathgroup at smc.vnet.net
Subject: [mg12992] [mg12937] Inconsistencies in pattern matching.


>This is an example taken from Ken Wagners Power Programming book:
>
>In[1]:={{x1,y1},{x2,y2},{x3,y3}}/.{x_,y_}->{x,Log[y]}
>
>Out[1]:={{x1,Log[y1]},{x2,Log[y2]},{x3,Log[y3]}}
>
>In[2]:={{x1,y1}}/.{x_,y_}->{x,Log[y]}
>
>Out[2]:={{x1,Log[y1]}}
>
>As long as the list of {x,y} data points has one point or greater than
>two points, it transforms as one would expect.  If there are  two {x,y}
>data points, it transforms differently.
>
>In[3]:={{x1,y1},{x2,y2}}/.{x_,y_}->{x,Log[y]}
>
>Out[3]:={{x1,y1},{Log[x2],Log[y2]}}
>
>
>Now, if x_ is seen to match {x1,y1} in example 3, then why doesn't it
>also match it in example number 1 and 2?  This behavior can be fixed
>with a /;Head[x]=!=List  pattern restricting rule, but that is not the
>point.  This seems grossly inconsistent to me.  Can anyone explain why
>it does this and/or justify that this is a good thing?
>



  • Prev by Date: Re: Find Max of "Concave" List
  • Next by Date: Re: Strange behavior of Sort
  • Previous by thread: Inconsistencies in pattern matching.
  • Next by thread: Re: Inconsistencies in pattern matching.