R help - Apr 2011 - get cells by the combination of their column and row names

Hi, 
I have a (correlation) matrix and I want to select a subset of its cells
depending on the combination of their column and row names.
This illustrates my problem: mtrx <- matrix(c(1,2,3,4,5,6,7,8,9), nrow=3,
ncol=3, dimnames = list(c('c132','c432', 'c233'), 
c('r132','r233', 'r432')))> mtrx     r132 r233
r432c132    1    4    7c432    2    5    8c233    3    6    9
At this point I want to compute the mean of the cells in which the column and
the row names share the same suffix (i.e., the cells mtrx[1,1], mtrx[2,3],
mtrx[3,2]). By suffix I mean the last three digits of the row/col name.
I need a generic code (and not vec<-c(mtrx[1,1], mtrx[2,3], mtrx[3,2]))
because I don't know in advance how many rows and columns I will have
(currently 118).
Thanks, Yoav 		 	   		  
	[[alternative HTML version deleted]]

Yoav -
    Here's one possibility:
> wh = outer(rownames(mtrx),colnames(mtrx),
+             function(x,y)substr(x,nchar(x)-2,nchar(x)) ==
substr(y,nchar(y)-2,nchar(y)))
> mtrx[wh]
[1] 1 6 8

If you knew that all of the row and column names were 4 characters long, it
would simplify to
> mtrx[outer(rownames(mtrx),colnames(mtrx),function(x,y)substr(x,2,4) ==
substr(y,2,4))]
Hope this helps.
 					- Phil Spector
 					 Statistical Computing Facility
 					 Department of Statistics
 					 UC Berkeley
 					 spector at stat.berkeley.edu



On Wed, 20 Apr 2011, yoav baranan wrote:
>
> Hi,
> I have a (correlation) matrix and I want to select a subset of its cells
depending on the combination of their column and row names.
> This illustrates my problem: mtrx <- matrix(c(1,2,3,4,5,6,7,8,9),
nrow=3, ncol=3, dimnames = list(c('c132','c432',
'c233'),  c('r132','r233', 'r432')))> mtrx   
r132 r233 r432c132    1    4    7c432    2    5    8c233    3    6    9
> At this point I want to compute the mean of the cells in which the column
and the row names share the same suffix (i.e., the cells mtrx[1,1], mtrx[2,3],
mtrx[3,2]). By suffix I mean the last three digits of the row/col name.
> I need a generic code (and not vec<-c(mtrx[1,1], mtrx[2,3], mtrx[3,2]))
because I don't know in advance how many rows and columns I will have
(currently 118).
> Thanks, Yoav
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> 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.
>

On Apr 20, 2011, at 3:52 PM, yoav baranan wrote:
>
> Hi,
> I have a (correlation) matrix and I want to select a subset of its  
> cells depending on the combination of their column and row names.
> This illustrates my problem: mtrx <- matrix(c(1,2,3,4,5,6,7,8,9),  
> nrow=3, ncol=3, dimnames = list(c('c132','c432',
'c233'),
> c('r132','r233', 'r432')))> mtrx     r132 r233
r432c132    1    4
> 7c432    2    5    8c233    3    6    9
> At this point I want to compute the mean of the cells in which the  
> column and the row names share the same suffix (i.e., the cells  
> mtrx[1,1], mtrx[2,3], mtrx[3,2]). By suffix I mean the last three  
> digits of the row/col name.
> I need a generic code (and not vec<-c(mtrx[1,1], mtrx[2,3],  
> mtrx[3,2])) because I don't know in advance how many rows and  
> columns I will have (currently 118).
 > mtrx[ sub("r|c", "", rep(rownames(mtrx),times=3) ) =  
sub("r|c", "", rep(colnames(mtrx), each=3))]
[1] 1 6 8

 > mean( mtrx[ sub("r|c", "", rep(rownames(mtrx),times=3)
) =               sub("r|c", "", rep(colnames(mtrx),
each=3))] )
[1] 5

Obviously you need to substitute NROW() and NCOL() vlues for the rep  
arguments but that should be simple.

> Thanks, Yoav 		 	   		
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> 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.
David Winsemius, MD
West Hartford, CT