@ Raul
> The J way of implementing merge might be /:~@,
I agree, but I want to avoid the use of /: :-)
I already tried your variant myself:
mergeSort1 =: merge & $: &>/ @ split ^: (1 < #) ^:_
but it is for arguments of length 3e3 10x slower (?), so I dropped that one.
M=. 3e3 ?...@#1e3
ts 'mergeSort1 M'
2.970816 14839488
ts 'mergeSort M'
0.324539 14807488
@ Viktor
> It's not always ridiculous
I meant compared to my implementation of 'merge'.
> a =. {...@] , (merge1 }.)
> merge1=: a...@.(>&{.)`,@.(*...@*&#)
> mergeSortV =: merge1 & $: &>/ @ split ` ] @. (1>:#)
mergeSortV 1 5 7 3 8 2 6 9 4
1 4 8 7 6 5 2 3 9
??
change to a nand function e.g. (or flip the outer gerund):
merge1=: a...@.(>&{.)`,@.(-...@*@*&#)
mergeSortV 1 5 7 3 8 2 6 9 4
1 2 3 4 5 6 7 8 9
M=. 3e3 ?...@#1e3
mergeSortV M
|stack error: merge1
| mergeSortV M
Return to not splitting up the merge verb:
merge1=: ({...@] , ($: }.))~` ({...@] , ($: }.)) @.(>&{.)`,@.(-...@*@*&#)
mergeSortV M
0 0 0 0 0 0 1 1 1 1 1 2 3 3 .....
FYI
Thank you for your replies.
--
Met vriendelijke groet,
=@@i
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
