It testifies to importance of good test data,
as provided in a separate message by Glenn.
Another correction--now obvious.
]r=: "."0;._2'101010101 100100100 100010001 000010000 010011100 '
1 0 1 0 1 0 1 0 1
1 0 0 1 0 0 1 0 0
1 0 0 0 1 0 0 0 1
0 0 0 0 1 0 0 0 0
0 1 0 0 1 1 1 0 0
]j=. i,I.s{~ i=.(i.>./)+/ s=.(= >./@,) (-.=i.#r)*.(+/ . *. |:) r
0 2
]k=. I.*./ j{r
0 4 8
Note (f/ . g |:) is the same as f/@:g"1/~
--- "Glenn M. Lewis" <[EMAIL PROTECTED]> wrote:
> Sorry for replying to my own post again, but after a nice shower (where
> I do my best thinking), I think I have almost figured out what Oleg wrote...
>
> First, he compares each row of the table with each other row with the
> following expression:
> ="1/~ t
> Next, he sums up the resulting planes into a single table:
> +/@:="1/~ t
> He then takes the identity matrix (using the size of the original
> table... which makes me wonder if this solution will work for a
> non-square table) and inverts it:
> (-.=i.1{$t)
> He multiplies this with the sums above to remove the main diagonal
> (which represents each row matching itself):
> (-.=i.1{$t)*.+/@:="1/~ t
> This now contains the number of times that each row matches other rows
> for each column. Then he marks the maximums in this table with ones and
> everything else with zeros:
> (= >./@,)(-.=i.1{$t)*.+/@:="1/~ t
> Now, things start to get a little fuzzy for me...
> This would sum the rows:
> +/s
> And this would perform the logical and of the columns:
> >./s
> But the expression (i.>./s) seems to generate the empty list, and i
> turns out to be zero.
> So I think I'm lost. Can anyone please shed some more light on this one
> for me?
> Thanks!
> -- Glenn
>
> Glenn M. Lewis wrote:
> > Wow!!! That is very impressive! I tried pulling apart the line (by
> > executing tiny bits at a time and displaying the line in boxed
> > format), and I still don't fully understand it yet, but I was able to
> > tweak it slightly to get it to produce a result... Here's my minor
> > tweak:
> >
> > t
> > 1 0 1 1 1 1 1 0
> > 0 1 1 1 1 0 1 1
> > 1 1 1 1 0 1 1 0
> > 1 1 1 1 0 1 1 1
> > 1 1 0 0 1 1 1 1
> > 1 0 1 1 1 1 1 1
> > 1 1 1 1 1 1 1 0
> > 0 1 0 1 1 0 1 1
> > ]j=. i,I.s{~i=.(i.>./)@:(+/) s=.(= >./@,)(-.=i.1{$t)*.+/@:="1/~t
> > 0 5 6
> > ]k=. I.*./ j{t
> > 0 2 3 4 5 6
> >
> > Thanks a bunch, Oleg! Hopefully before my next birthday I will have
> > figured out what you just wrote. :-)
> > -- Glenn
> >
____________________________________________________________________________________
Bored stiff? Loosen up...
Download and play hundreds of games for free on Yahoo! Games.
http://games.yahoo.com/games/front
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
