| rowMedians {Biobase} | R Documentation |
Calculates the median for each row in a matrix.
rowMedians(imat, na.rm=FALSE)
imat |
A numeric matrix. |
na.rm |
If TRUE, NAs are excluded before calculating the medians, otherwise not. |
... |
Not use. |
Returns a double vector of length equal to number of rows in x.
Missing values are excluded before calculating the medians.
This implementation is optimized for speed and memory to calculate.
As the example shows, this implementation is roughly 3-10 times faster
than using apply(x, MARGIN=1, FUN=medians).
As the example might show, the rowQ() does not (have to)
handle missing values, and is therefore in some cases faster.
Henrik Bengtsson
See rowMeans() in colSums().
set.seed(1) x <- rnorm(n=234*543) x[sample(1:length(x), size=0.1*length(x))] <- NA dim(x) <- c(234,543) y1 <- rowMedians(x, na.rm=TRUE) y2 <- apply(x, MARGIN=1, FUN=median, na.rm=TRUE) stopifnot(all.equal(y1, y2)) x <- cbind(x1=3, x2=c(4:1, 2:5)) stopifnot(all.equal(rowMeans(x), rowMedians(x)))