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
