[R] apply & incompatible dimensions error

Bernzweig, Bruce (Consultant) bbernzwe at bear.com
Tue Jul 24 17:47:06 CEST 2007


Thanks for the explanation.

As for the rows/columns thing, the data I receive is given to me in that
way.  I currently read it in using read.csv.  Is there a function I
should look at that can take that and transpose it or should I just
process the data first outside of R?

Thanks,

- Bruce

-----Original Message-----
From: Gabor Grothendieck [mailto:ggrothendieck at gmail.com] 
Sent: Tuesday, July 24, 2007 11:43 AM
To: Bernzweig, Bruce (Consultant)
Cc: r-help at stat.math.ethz.ch
Subject: Re: [R] apply & incompatible dimensions error

Then try this:

cor(t(mat1), t(mat2))

Also note

1. the above implies that mat1 and mat2 must have the same
number of columns since if x and y are vectors cor(x,y) only makes
sense if they have the same length.

2. the usual convention is that variables are stored as columns
andt that rows correspond to cases so typically you would have
(in terms of your mat1 and mat2):

Mat1 <- t(mat1)
Mat2 <- t(mat2)

and then use Mat1 and Mat2, e.g. cor(Mat1, Mat2)



On 7/24/07, Bernzweig, Bruce (Consultant) <bbernzwe at bear.com> wrote:
> Thanks Gabor!
>
> You state that my apply is taking rows of mat1 with columns of mat2.
>
> Is this because I have the y=mat2 parameter?
>
> > apply(mat1, 1, f, y=mat2)
>
> Actually, what I would like is to run the correlations on a row
against
> row basis:  mat1 row1 x mat2 row1, etc.
>
> Thanks again,
>
> - Bruce
>
>
> -----Original Message-----
> From: Gabor Grothendieck [mailto:ggrothendieck at gmail.com]
> Sent: Tuesday, July 24, 2007 11:31 AM
> To: Bernzweig, Bruce (Consultant)
> Cc: r-help at stat.math.ethz.ch
> Subject: Re: [R] apply & incompatible dimensions error
>
> Your apply is trying to take the correlations of the rows of mat1 with
> the
> columns of mat2 which, of course, does not work if they have different
> numbers of columns. I think you mean to take the correlations of the
> columns
> of mat1 with the columns of mat2.  For example, to take the
correlations
> of the 5 columns of mat1 with the first 4 columns of mat2 try:
>
> > cor(mat1, mat2[,1:4])
>            col1       col2       col3       col4
> col1 -0.34624254 -0.2669519 -0.2705077  0.2183249
> col2 -0.26553255 -0.2687643 -0.0865895  0.1819025
> col3  0.19474613 -0.2334986  0.1746522  0.2326915
> col4  0.09328338  0.5117784  0.2413143 -0.3374916
> col5  0.27519716  0.1605331 -0.4057137  0.3282105
>
>
> On 7/24/07, Bernzweig, Bruce (Consultant) <bbernzwe at bear.com> wrote:
> > Hi,
> >
> > I've created the following two matrices (mat1 and mat2) and a
function
> > (f) to calculate the correlations between the two on a row by row
> basis.
> >
> >        mat1 <- matrix(sample(1:500,50), ncol = 5,
> >                dimnames=list(paste("row", 1:10, sep=""),
> >                paste("col", 1:5, sep="")))
> >
> >        mat2 <- matrix(sample(501:1000,50), ncol = 5,
> >                dimnames=list(paste("row", 1:10, sep=""),
> >                paste("col", 1:5, sep="")))
> >
> >        f <- function(x,y) cor(x,y)
> >
> > When the matrices are squares (# rows = # columns) I have no
problems.
> >
> > However, when they are not (as in the example above with 5 columns
and
> > 10 rows), I get the following error:
> >
> > > apply(mat1, 1, f, y=mat2)
> > Error in cor(x, y, na.method, method == "kendall") :
> >        incompatible dimensions
> >
> > Any help would be appreciated.  Thanks!
> >
> > - Bruce
> >
> >
> >
> >
**********************************************************************
> > Please be aware that, notwithstanding the fact that the
> pers...{{dropped}}
> >
> > ______________________________________________
> > R-help at stat.math.ethz.ch 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.
> >
>
>
>
> **********************************************************************
> Please be aware that, notwithstanding the fact that the person sending
> this communication has an address in Bear Stearns' e-mail system, this
> person is not an employee, agent or representative of Bear Stearns.
> Accordingly, this person has no power or authority to represent, make
> any recommendation, solicitation, offer or statements or disclose
> information on behalf of or in any way bind Bear Stearns or any of its
> affiliates.
> **********************************************************************
>



**********************************************************************
Please be aware that, notwithstanding the fact that the pers...{{dropped}}



More information about the R-help mailing list