Thanks to all for the help. I have done the following tests and so far
zsban2 comes out as the best solution.
raul1 =: ] {~ ~: >./\@:* [EMAIL PROTECTED]@]
zsban1 =: (~:(#{~[:<:+/[EMAIL PROTECTED])])
zsban2 =: (~:(#;.1##)]) NB. Mike also?
oleg1 =: [^:(0=])/\
brand1 =: ( ((] {~ <:@:(I.@:=))`(I.@:=)`])} ) ^:_
y =: 0 0 1 0 2 0 0 3 0 0 0 4 0 0 0 0
0 raul1 y
0 0 1 1 2 2 2 3 3 3 3 4 4 4 4 4
0 zsban1 y
4 4 1 1 2 2 2 3 3 3 3 4 4 4 4 4
0 zsban2 y
1 1 2 2 2 3 3 3 3 4 4 4 4 4
oleg1 y
0 0 1 1 2 2 2 3 3 3 3 4 4 4 4 4
0 brand1 y
4 4 1 1 2 2 2 3 3 3 3 4 4 4 4 4
NB. time
10000 (6!:2) '0 raul1 y'
4.19281e_6
10000 (6!:2) '0 zsban1 y'
5.02911e_6
10000 (6!:2) '0 zsban2 y'
3.78488e_6
10000 (6!:2) 'oleg1 y'
0.000147781
10000 (6!:2) '0 brand1 y'
2.25551e_5
NB. space
(7!:2) '0 raul1 y'
1280
(7!:2) '0 zsban1 y'
1216
(7!:2) '0 zsban2 y'
1280
(7!:2) 'oleg1 y'
2432
(7!:2) '0 brand1 y'
5824
NB. Summary
NB. zsban2 is fastest and fills leading zeroes with
NB. the first non-zero which is desirable.
NB. It also does well in terms of memory used.
On Thu, Oct 30, 2008 at 6:07 AM, Oleg Kobchenko <[EMAIL PROTECTED]> wrote:
>> From: Matthew Brand <[EMAIL PROTECTED]>
>
>>
>> Hi All,
>>
>> Is there a tacit way to copy forward the number which occurs before
>> the first zero in any sequence of zeroes into the positions of those
>> zeroes?
>>
>> E.g. Is there a tacit f which does this:
>> 0 f 1 0 2 3 4 5 0 0 0 4 3 2 4
>> 1 1 2 3 4 5 5 5 5 4 3 2 4
>
> Not strictly dyadic, but here is a simple tacit idea:
>
> [^:(0=])/\ 1 0 2 3 4 5 0 0 0 4 3 2 4
> 1 1 2 3 4 5 5 5 5 4 3 2 4
>
>
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
