[R] Error in colMeans ... what's wrong with my data?

[Prof Brian Ripley]

On Fri, 1 Apr 2005, Hartmut Weinrebe wrote:

> Having searched and searched I still haven't found what's the problem with my data (I've attached the relevant file).
> Every time I tried to use the CANCOR-Function I got error messages.
> So I turned to check my two sets of variables separately by using the "colMeans"-Function. With one set no problem .. but with the other.
>
> My input :
>
>> struktur <- read.delim("struktur.csv", header=TRUE, sep = ";")
>> colMeans(struktur)
>
> And the resulting error:
>
> Error in colMeans(x, n, prod(dn), na.rm) :
>        `x' must be numeric
>
>
> Can anyone give me a hint concerning the mistake I made?

You applied colMeans to a data frame: the help page says

        x: an array of two or more dimensions, containing numeric,
           complex, integer or logical values, or a numeric data frame.

One or more columns of `struktur' are not numeric.  (Note that whereas an 
array can be logical, a data frame must be wholly numeric.)

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595