[R] Matrix formatting

[Adaikalavan Ramasamy]

I do something like this as part of a missing value plot where I 'flush'
the rows with most missing spots to the bottom and the columns with the
most missing values to the right. In this manner I can visually see how
much data I will retain (by looking at the area) if I decide to truncate
rows/columns by missing value criterion. You might find this code useful
in rearranging your matrix:

na.mat <- 1*is.na(data)                           
spots.na.per.row <- rowSums(na.mat)/ncol(data)    # calculates the
percentage missing by row
spots.na.per.column <- colSums(na.mat)/nrow(data)

data.re <- data[ order(spots.na.per.row), order(spots.na.per.column) ]



-----Original Message-----
From: Jerome Asselin [mailto:jerome at hivnet.ubc.ca] 
Sent: Thursday, February 13, 2003 5:06 AM
To: Tim Sharac; r-help at stat.math.ethz.ch
Subject: Re: [R] Matrix formatting



Would this do what you want?

Cheers,
Jerome

> foo.dat <- cbind(c(NA, 1, 2, 3, 4, 5), c(NA, NA, 0, 10 ,20, 30))
> apply(foo.dat,2,function(x) x[order(as.logical(x))])
     [,1] [,2]
[1,]    1    0
[2,]    2   10
[3,]    3   20
[4,]    4   30
[5,]    5   NA
[6,]   NA   NA


On Wednesday 12 February 2003 12:42, Tim Sharac wrote:
> Content-Length: 918
> Status: R
> X-Status: N
>
> Hi R-users:
>
> I have a data formatting question. I have a data set that looks 
> something like this:
>
> foo.dat <- cbind(c(NA, 1, 2, 3, 4, 5), c(NA, NA, 0, 10 ,20, 30))
>
> What I have:
>
>      [,1] [,2]
> [1,]   NA   NA
> [2,]    1   NA
> [3,]    2    0
> [4,]    3   10
> [5,]    4   20
> [6,]    5   30
>
>
> I want to line up the columns by the first value that is not NA. Like 
> so:
>
>      [,1] [,2]
> [1,]    1   0
> [2,]    2   10
> [3,]    3   20
> [4,]    4   30
> [5,]    5   NA
> [6,]    NA  NA
>
> Question is: Is there an elegant way to do this without a for loop?
>
> I tried doing this with na.omit and na.exclude without success.
>
> The real data is many hundreds of columns and many thousands of rows.
>
> Thanks in advance, Tim
>
> Sign up for Internet Service under $10 dollars a month, at 
> http://isp.BlueLight.com
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list 
> http://www.stat.math.ethz.ch/mailman/listinfo/r-help

______________________________________________
R-help at stat.math.ethz.ch mailing list
http://www.stat.math.ethz.ch/mailman/listinfo/r-help