[R] Problem with by(... , median)
figurski at mail.med.upenn.edu
Tue Jun 10 15:18:22 CEST 2008
I am new to R, I have been using SAS for a while. Not surprisingly, I
find R much better in graphics, which is publication ready right away.
Recently, I have been trying to calculate some basic statistics using R.
I have a dataset of multiple rows per subject. For example:
Subject Date Factor1 Factor2 Factor3
P1 0.5 1 1 3
P1 1 3 2 5
P1 2 3 5 NA
P2 0.5 1 6 4
P2 1 2 NA 7
P2 2 3 1 2
I wanted to get the 'by subject' statistics of the factors. I used the
following code: a=by(dat1, dat1$Subject, mean, na.rm=TRUE) - this is for
However, when I try that with a 'median', I get an error message. It is
strange to me, because with the 'summary' option, I get the median
values for all factors separately, among other information. What I need
is just a median, because I later need it for plotting.
I have read old posts on the archive, and I agree that the action of
'median' is according to the documentation. However, I think this is a
serious inconvenience, and it should be changed to work in the same way
as mean, summary or sd options.
Nevertheless, accepting the reality, I have a question to all: how to
obtain the by-subject medians of factors in such dataset?
Thanks in advance,
Michal J. Figurski
HUP, Pathology & Laboratory Medicine
Xenobiotics Toxicokinetics Research Laboratory
3400 Spruce St. 7 Maloney
Philadelphia, PA 19104
More information about the R-help