[R-sig-eco] doubt in generating loops with For

[Thiago A. Pires]

Hi Vít,

Thank you for your quick response. Indeed, I did not put all the code,
because I thought it would mess up more than help. But I'll put here the
full code. I realized the issue of columns, I had come to this conclusion
also. However, I would like to generate a "random" matrix of  community
based in a "real" community  in every loop, and likewise, generating a
diversity measure of these communities, also in each loop. If you can help
me with this question , I'd appreciate.

Code:

require (vegan)

comm<-t(data.frame(sp1=rpois(10, 1), sp2=rpois(10,1), sp3=rpois(10,1),
                       sp4=rpois(10,1), sp5=rpois(10,1)))

traits<-t(data.frame(sp1=rnorm(5,1.2,.2),sp2=rnorm(5,2,.5),
                     sp3=rnorm(5,3,.3), sp4=rnorm(5,0.5,.2),
sp5=rnorm(5,.9,.6)))

dist<-vegdist(traits,"gower")
comm<-as.matrix(comm)
func.sim<-data.frame(rep(NA, 1000))

 for (z in 1:1000)
  {
  comm1<-apply(comm,2,sample)
  func.sim[z,]<- (t(comm[, z]) %*% (as.matrix(dist)^2) %*% comm[,
z])/2/(sum(comm[, z])^2)
  }

Error in comm[ ,z] : subscript out of bounds





Best regards,

Thiago



_____________________________________________________________

*Thiago A. Pires*
Ph.D. student in Ecology -  Instituto de Biologia - UNICAMP
Collaborator Researcher - Laboratório de Ecologia e Sistemática - UNIFESP


Email: pires.thiagoaugusto at gmail.com / Skype: thiago.augusto.pires
Home Page:
http://thiagoaugustopires.wix.com/naturevisions
CV Lattes:
http://buscatextual.cnpq.br/buscatextual/visualizacv.do?id=K4298681Z0
_____________________________________________________





2015-04-16 4:40 GMT-03:00 Vít Syrovátka <syrovat at sci.muni.cz>:

> Hi Thiago,
> I am not familiar with the diversity measure you are trying to calculate
> and your example code is not complete - one can not simulate your comm and
> dist objects.
> But at the first sight one can see that in your loop you are trying to
> select columns 1, 2, 3,... 1000 from the comm matrix, which probably has
> only 10 columns (if sites are in columns and species in rows, contrary to
> how such data are traditionally stored).
>
> for(z in 1:10)
> instead of
> for(z in 1:1000)
> should do the job.
>
> Vit
>
> Dne 2015-04-15 23:24, Thiago A. Pires napsal:
>
>> Hi guys,
>>
>> I have a question when trying to generate 1000 measures with "for"
>> function. For example, I have 10 locations with 5 species , and I want to
>> apply a diversity measure on each site. But when I try to run one is with
>> my function, I can not, I get an error message: "Error in com[, z] :
>> subscript out of bounds", and I think it should be a simple mistake, but
>> I'm not seeing. If you guys can help with it, I would appreciate. Right
>> below is an example for my code:
>>
>> comm= a community matrix with 10 sites and 5 species
>> dist= a pairwise distance-based matrix of species
>> ​
>> ​ for (z in 1:1000)
>>   {
>>
>>   func.sim[z,]<- (t(comm[, z]) %*% (as.matrix(dist)^2) %*%
>>                  comm[, z])/2/(sum(comm[, z])^2)​
>>
>>   }
>>
>>
>> ​Error in comm[, z] : subscript out of bounds​
>>
>>
>>
>>
>>
>>
>> ​Best regards,
>>
>> Thiago.​
>>
>>
>> _____________________________________________________________
>>
>> *Thiago A. Pires*
>> Ph.D. student in Ecology -  Instituto de Biologia - UNICAMP
>> Collaborator Researcher - Laboratório de Ecologia e Sistemática - UNIFESP
>>
>>
>> Email: pires.thiagoaugusto at gmail.com / Skype: thiago.augusto.pires
>> Home Page:
>> http://thiagoaugustopires.wix.com/naturevisions
>> CV Lattes:
>> http://buscatextual.cnpq.br/buscatextual/visualizacv.do?id=K4298681Z0
>> _____________________________________________________
>>
>>         [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> R-sig-ecology mailing list
>> R-sig-ecology at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
>>
>

	[[alternative HTML version deleted]]