[R] colnames from submatrix?

[Liaw, Andy]

If the columns you want to exclude can only contain all 0s (and not some
other values such as 1s), you can do something like:

mat[, colSums(mat != 0) > 0]

to extract the non-zero columns of the matrix.  "mat != 0" gives you a
matrix of TRUEs and FALSEs, and colSums() of this matrix tells you how many
non-zero elements are in the columns.  You just want columns with at least
one non-zero entry.

HTH,
Andy

> -----Original Message-----
> From: Jeff D. Hamann [mailto:jeff_hamann at hamanndonald.com] 
> Sent: Monday, October 13, 2003 12:22 PM
> To: r-help at stat.math.ethz.ch
> Subject: [R] colnames from submatrix?
> 
> 
> Hi R-Wizards:
> 
> I've looking through the R docs and have yet to find what I'm 
> looking for and have tried a few intermediate steps to now 
> avail yet and rather than spend another few hours looking for 
> the solution, I figured I would post a message.
> 
> I have a matrix (actually a set of them) that I want to pull 
> all the names of the non-zero columns into a vector/list for 
> further processing:
> 
>              d0            d1            d2 s0 s1 s2 s3
>  [1,] -4.4721360 -447.29878969 -436.18978037  0  0  0  0
>  [2,]  0.2236068  -25.83121172  -29.21571675  0  0  0  0
>  [3,]  0.2236068    0.13009088   42.49338060  0  0  0  0
>  [4,]  0.2236068    0.17155235    0.04891280  0  0  0  0
>  [5,]  0.2236068   -0.08027479   -0.14749423  0  0  0  0
>  [6,]  0.2236068   -0.02394758   -0.12845820  0  0  0  0
>  [7,]  0.2236068    0.03838012   -0.15270556  0  0  0  0
>  [8,]  0.2236068    0.18150155   -0.17073267  0  0  0  0
>  [9,]  0.2236068   -0.14047328   -0.11013806  0  0  0  0
> [10,]  0.2236068   -0.34247697   -0.05627832  0  0  0  0
> [11,]  0.2236068   -0.27058719    0.31379811  0  0  0  0
> [12,]  0.2236068   -0.04930450    0.41093718  0  0  0  0
> [13,]  0.2236068    0.10968920    0.32996356  0  0  0  0
> [14,]  0.2236068   -0.05104658    0.08743581  0  0  0  0
> [15,]  0.2236068   -0.19184524   -0.29548841  0  0  0  0
> [16,]  0.2236068   -0.06285400   -0.26123732  0  0  0  0
> [17,]  0.2236068   -0.52558878   -0.34821174  0  0  0  0
> [18,]  0.2236068    0.15258305   -0.10895070  0  0  0  0
> [19,]  0.2236068    0.22044668   -0.21442723  0  0  0  0
> [20,]  0.2236068    0.51934865   -0.41193918  0  0  0  0
> 
> 
> such that I only end up with a list of names (d0,d1,d2). I've 
> been using the qr( x )$rank to get the number of non-zero 
> columns, but I'm not sure how to only return the names of the 
> non-zero columns as
> 
> > colnames( qr( attr( eval( nlsystemols$eq[[1]]$deriv ), "gradient" ) 
> > )$qr )
> 
> will return
> 
> > [1] "d0" "d1" "d2" "s0" "s1" "s2" "s3"
> 
> Is/Are there one of those great S or R function/shortcuts 
> that will return a
> submatrix(?) or a partioned matrix such that when I call 
> colnames() I get:
> 
> > [1] "d0" "d1" "d2"
> 
> Thanks,
> Jeff.
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list 
> https://www.stat.math.ethz.ch/mailman/listinfo> /r-help
>