That should be in the phrases rather than what I used!
Ric's definition also  works but not with a 5 element list
   y=:2 3 6 8 9
   = y
1 0 0 0 0
0 1 0 0 0
0 0 1 0 0
0 0 0 1 0
0 0 0 0 1
   =/~y
1 0 0 0 0
0 1 0 0 0
0 0 1 0 0
0 0 0 1 0
0 0 0 0 1

Thank you both - so many ways to skin a cat!

Don
On 19/11/2012 9:01 PM, Bo Jacoby wrote:
The unit matrix is
    =/~i.5
1 0 0 0 0
0 1 0 0 0
0 0 1 0 0
0 0 0 1 0
0 0 0 0 1

- Bo





________________________________
Fra: Don & Cathy Kelly <d...@shaw.ca>
Til: programm...@jsoftware.com
Sendt: 0:47 tirsdag den 20. november 2012
Emne: Re: [Jprogramming] Verb definition + inadvertent train formation.

No - I should have made it clear that the actual y is a list of n
numbers and I need to add this list to the diagonal of a matrix.
Since this will (of necessity) be within a loop and each time both the
list and matrix will be modified (except for size),
I have come to the conclusion that I should make a single unit matrix
outside the iteration and use this each time.
    ] u=:=@i.@#i.5

1 0 0 0 0
0 1 0 0 0
0 0 1 0 0
0 0 0 1 0
0 0 0 0 1

y=: 2 3 5 4 10

     ]u *y
2 0 0 0  0
0 3 0 0  0
0 0 5 0  0
0 0 0 4  0
0 0 0 0 10
repeating with different y values as needed.

The algorithm for u is taken from the "phrases" section of the
vocabulary-maybe you have a simpler one?-
Thanks

Don


On 18/11/2012 10:33 PM, bob therriault wrote:
Hi Don,

Are you looking for

y=:5
      c=:(*=)&i.
      c y
0 0 0 0 0
0 1 0 0 0
0 0 2 0 0
0 0 0 3 0
0 0 0 0 4
Cheers, bob

On 2012-11-18, at 7:13 PM, Don & Cathy Kelly wrote:

OOPs' I meant y=:i.5 5

Don
On 18/11/2012 7:04 PM, Don & Cathy Kelly wrote:
That gives the diagonal nicely  but the problem that I didn't express clearly 
is that I want a matrix as follows
     y=:i.5

     ]result =: (* =@i.@#) y
0 0 0 0 0
0 1 0 0 0
0 0 2 0 0
0 0 0 3 0
0 0 0 0 4
which II can get by
c=:* =@i.@#
     c y
0 0 0 0 0
0 1 0 0 0
0 0 2 0 0
0 0 0 3 0
0 0 0 0 4

For my purposes I can use a base nth order unit matrix by the phrase as given 
in phrases as  m=:=@i.@# y
and then deal with m* y1 , m* y2 etc which suits my purpose which is to form a 
specific Jacobean used for power system load flows.
I have handled this problem in APL  and am trying to replicate the coding more 
directly in J-mainly as an exploration of J for my own satisfaction.

Thank you

Don Kelly
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm



----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm


----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to