[R] how to pass additional parameters to a function called in tapply?

Wolfgang Waser wolfgang.waser at utu.fi
Mon Sep 1 19:41:38 CEST 2008


Hi all,

the following problem is still beyond my R-knowledge:

I have one data vector containing the signal from 4 channels that are measured 
subsequently and in repeating cycles (with one factor vector for cycle and 
one for channel identification).

To extract the mean of each channel during each cycle tapply is the method of 
choice. However, I cannot use the whole measuring period for each channel as 
marked in the factor vector since the hardware switching between channels 
causes some transients in the signal which should not be included when 
calculating the mean.


Instead of using "mean" as function within tapply, I used the following set of 
functions:


start_end.f <- function(x) { c(length(x)*(1-part),length(x)) }


to define the part of the data to use (e.g. part=0.5 the last 50% of the total 
period),

#(part has so far been set universally once for the whole analysis, but see 
below)#


avg.f <- function(x) {
  range <- start_end.f(x)
  mean(x[range[1]:range[2]])
}


and then


tapply(data,factor-vectors,avg.f)


to extract the mean of a subsample of the total measuring period of each 
channel.


The problem now is that I have several signals and different 'parts' should be 
analysed for each, so functions have to be changed to

start_end.f <- function(x,part) { ... }

avg.f <- function(x,part)
  range <- start_end.f(x,part)
  mean(x[range[1]:range[2]])
}


However, this causes an error


Error in FUN(X[[1L]], ...) : argument "part" is missing, with no default <- 


since tapply seems to be unable to pass more than x to the function called



Is there a way to pass variables to the function called in tapply other than 
the dataset extracted by tapply?



Thanks for any help

Wolfgang



More information about the R-help mailing list