R help - Jun 2012 - select subrows based on a specific column in a matrix

Hi all,

I have a matrix with 10000 rows and 10 columns. The last columns contains
another identifiers but the values are not uniques so that I want to generate
another matrix with rows with unique values in the last column.

If I did

tmp<-unique(my_mat$col10) 

this will give me 8560 unique entries so the ideal matrix will be 8560X10
columns now then.

I tried

sub_mat<-my_mat[tmp,]

but it generated weird results with many "NA" values and the order was
not changed. The original matrix was ranked from top so I don't want to lose
the order too.

For the similar problem, I have used "match" function and do some
manipulate to identify the index of the first appearance of each value but is
there any better and neat way to achieve the same function?

Thanks,

Seungyeul Yoo

Postdoc Fellow,
Institute of Genomics and Multiscale Biology
Department of Genetics and Genomic Sciences
Mount Sinai School of Medicine

Hello,

You should post a data example, like the posting guide says. If your 
dataset is large, use something like

dput(head(dat, 20))   # paste the output of this in your post.


where 'dat' is your dataset.
Now, try


# make up some data
set.seed(12)
dat <- matrix(c(sort(rnorm(10)), sample(letters[1:4], 10, TRUE)), ncol=2)
colnames(dat) <- c("A", "col10")
dat

# this does it
ix <- as.logical(ave(seq_len(nrow(dat)), dat[, "col10"],
         FUN=function(x) ifelse(x == min(x), TRUE, FALSE)))
dat[ix, ]  # rows 1, 2, 4, 6


Hope this helps,

Rui Barradas

Em 07-06-2012 17:07, Seungyeul Yoo escreveu:
> Hi all,
>
> I have a matrix with 10000 rows and 10 columns. The last columns contains
another identifiers but the values are not uniques so that I want to generate
another matrix with rows with unique values in the last column.
>
> If I did
>
> tmp<-unique(my_mat$col10)
>
> this will give me 8560 unique entries so the ideal matrix will be 8560X10
columns now then.
>
> I tried
>
> sub_mat<-my_mat[tmp,]
>
> but it generated weird results with many "NA" values and the
order was not changed. The original matrix was ranked from top so I don't
want to lose the order too.
>
> For the similar problem, I have used "match" function and do some
manipulate to identify the index of the first appearance of each value but is
there any better and neat way to achieve the same function?
>
> Thanks,
>
> Seungyeul Yoo
>
> Postdoc Fellow,
> Institute of Genomics and Multiscale Biology
> Department of Genetics and Genomic Sciences
> Mount Sinai School of Medicine
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>