Re: part assigned sequence behavior puzzling
- To: mathgroup at smc.vnet.net
- Subject: [mg105393] Re: part assigned sequence behavior puzzling
- From: "Sjoerd C. de Vries" <sjoerd.c.devries at gmail.com>
- Date: Tue, 1 Dec 2009 04:13:03 -0500 (EST)
- References: <hekujm$96g$1@smc.vnet.net> <hf09b2$d1s$1@smc.vnet.net>
Hi Daniel,
Actually, in all my years with Mathematica I haven't had a single
occasion in which Sequence would used in this way, so perhaps the
whole discussion is moot.
> (1) What is the expected/intended behavior of
>
> In[87]:= ss = {a, b, c, d}; Do[ss[[j]] = Sequence[j, j + 1], {j,
> Length[ss]}]
Well, I know now that it is supposed to be {1, 2, 2, 3, 3, 4, 4, 5},
but my intuition would say {1, 2, 3, 4, 5, b, c, d}.
Look at this sequence (sic) of instructions (and execute it as a
whole):
s = {1, 2, 3}
s[[1]] = Sequence[4, 5]
s[[1]]
s[[2]]
s[[2]] = DifferentValueForIndex2
s[[2]]
s
The first time you ask for the contents of s[[2]] it is 5 (and this is
the only 5 you put in the list), then you assign something else to s
[[2]], so any average programmer would think the 5 is gone out of the
list now. But, surprise, there it still is... big magic
I'd say if you can check and use the value at index 2 for
calculations, you should be able to set the value at index 2, but you
can't in this case.
I wonder how many of your colleagues, when shown the subsequence of
s[[2]]
s[[2]] = DifferentValueForIndex2
s[[2]]
without showing them the first part, will be utterly amazed, and won't
know what's causing this.
Cheers -- Sjoerd
On Nov 30, 1:12 pm, d... at wolfram.com wrote:
> > I'd say that this is not a wise design decision. IMHO, a statement
> > like tmp=tmp should have no effect on tmp, whereas with this
> > SequenceHold situation it has.
>
> > Cheers -- Sjoerd
>
> Perhaps so. I view it, paraphrasing Churchill, as the worst possible
> design, except for all the rest.
>
> In the alternative scenario, Sequence[...] things would (I guess) cause
> immediate evaluation of the left hand side. This has some bad consequence=
s
> and maybe also raises a few questions.
>
> Bad consequences:
>
> (1) It changes assignment semantics.
> (2) It means that deeply nested expressions cannot readily be built
> because complexity of adding a level becomes worse (maybe this can largel=
y
> be short-circuited, I'm not sure).
>
> Questions:
>
> (1) What is the expected/intended behavior of
>
> In[87]:= ss = {a, b, c, d}; Do[ss[[j]] = Sequence[j, j + 1], {j, Le=
ngth[ss]}]
> ?
>
> (2) What about
> ss2 = Table[Unevaluated[Sequence[i, i^2]], {i, 4}]
> ?
>
> Maybe that second one is obvious, and no different from current behavior.
>
> All the same, I think it would not be good to have a semantics wherein th=
e
> assignment in the Do loop causes evaluation of the lhs in each iteration.
>
> Daniel Lichtblau
> Wolfram Research
>
> > On Nov 26, 6:01 am, Bob Hanlon <hanl... at cox.net> wrote:
> >> As stated in the documentation for sequence: assignment and replacemen=
t
> >> f=
> > unctions have attribute SequenceHold
>
> >> Attributes[Set]
>
> >> {HoldFirst,Protected,SequenceHold}
>
> >> tmp = Range[15]
> >> tmp[[7]] = Sequence @@ Range[2];
> >> tmp[[7]] = Sequence @@ Range[2];
> >> tmp[[7]] = Sequence @@ Range[2];
> >> tmp
>
> >> {1,2,3,4,5,6,7,8,9,10,11,12,13,14,15}
>
> >> {1,2,3,4,5,6,1,2,8,9,10,11,12,13,14,15}
>
> >> tmp = Range[15]
> >> tmp[[7]] = Sequence @@ Range[2];
> >> tmp = tmp;
> >> tmp[[7]] = Sequence @@ Range[2];
> >> tmp = tmp;
> >> tmp[[7]] = Sequence @@ Range[2];
> >> tmp
>
> >> {1,2,3,4,5,6,7,8,9,10,11,12,13,14,15}
>
> >> {1,2,3,4,5,6,1,2,2,2,8,9,10,11,12,13,14,15}
>
> >> Bob Hanlon
>
> >> ---- mkr <mileskra... at gmail.com> wrote:
>
> >> =============
> >> I am puzzled by the following behavior:
>
> >> tmp = Range[15]
> >> tmp[[7]] = Sequence @@ Range[2];
> >> tmp[[7]] = Sequence @@ Range[2];
> >> tmp[[7]] = Sequence @@ Range[2];
> >> tmp
>
> >> yields
>
> >> {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15}
> >> {1, 2, 3, 4, 5, 6, 1, 2, 8, 9, 10, 11, 12, 13, 14, 15}
>
> >> I would have expected the repeated assignment to have a repeated
> >> effect, thus obtaining
>
> >> {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15}
> >> {1, 2, 3, 4, 5, 6, 1, 2, 2, 2, 8, 9, 10, 11, 12, 13, 14, 15}
>
> >> Where/why am I wrong?