Re: split again
- To: mathgroup at smc.vnet.net
- Subject: [mg73667] Re: split again
- From: "Ray Koopman" <koopman at sfu.ca>
- Date: Sat, 24 Feb 2007 02:16:04 -0500 (EST)
- References: <ermd8n$hmd$1@smc.vnet.net>
On Feb 23, 1:48 am, Arkadiusz.Ma... at gmail.com wrote:
> Hi,
>
> Thank you very much for your help in my provious post. Now, consider
> please a list
>
> z=Table[Random[],{1000}];
>
> zSplit=Split[z,#1<=0.7 === #2>=0.7]; (bulit thanks to your
> help)
>
> I want to pick the first sublist of zSplit that consists of elements
> <= 0.7 and whose length is greater than a certain number (say 5). I
> think that a good candidate would be using Cases and variations of _ ,
> but I don't know how.
>
> What I want to do (and what my both posts are about) is to find the
> first sublist of z with all elements less than A and the length of
> this sublist must be greater than B. Maybe there exists better
> solution than to Split z in advance since what I need to do in my next
> step is to find ONLY the FIRST sublist of splitted z.
>
> Thanks again,
>
> Arek
SeedRandom[2]; z = Table[Random[],{20}]
{0.238705,0.844529,0.473928,0.421515,0.512692,
0.0228035,0.529257,0.0507012,0.74907,0.632356,
0.0725195,0.251276,0.902785,0.672727,0.419794,
0.512577,0.404555,0.00984124,0.168643,0.50482}
A = .7; B = 5;
n = Flatten[Position[isplit = Split@UnitStep[z - A],
_?(Length@# > B && First@# == 0&), {1}, 1]/.{}->{0}][[1]]
3
If no run exists with the desired properties then n will be 0.
Otherwise, n is the position in isplit of the desired run, and
Length@Flatten@Take[isplit,n-1] + {1, Length@isplit[[n]]}
{3,8}
gives the first & last positions of the run in z.