Hi all!
It might seem strange that you can't easily address the columns with Rank.
Alignment has a similar irregularity:
10 20 +i. 2 3
10 11 12
23 24 25
(i. 2 3)+10 20 30
|length error
| (i.2 3) +10 20 30
(i. 2 3)+10 20
10 11 12
23 24 25
Common first parts of the argument shapes are aligned. The smaller
argument is then replicated over the remaining dimensions of the larger
argument.
There is no alignment operator which aligns the last parts of the shapes
and replicates over the remaining first dimensions.
What is even more strange is that problems like these are never resolved.
A problem you had in IBM APL1 1980 usually remains in J today.
However, for both these problems, I think they were created with J. In
APL1 I think you wouldn't get the length error above. You could also
address columns.
I think J is better in both indexing and that you get the same result if
you change the order of the arguments with alignment. However, I think
there could be a way to address the columns. There could be a way to get
alignment along the last dimensions.
Happily, these problems often cancel each other. What you can't do with
alignment you can do with Rank.
(i. 2 3)+"1 [10 20 30
10 21 32
13 24 35
Cheers,
Erling Hellenäs
Den 2017-11-27 kl. 04:21, skrev 'Skip Cave' via Programming:
Combination of 6 items taken 2 at a time:
comb=: ((= +/"1) |.@:I.@# ]) #:@i.@(2&^)
|: 2 comb 6
0 0 0 0 0 1 1 1 1 2 2 2 3 3 4
1 2 3 4 5 2 3 4 5 3 4 5 4 5 5
How can I box each column with the vertical pair of numbers in each box?
Skip
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
