[R] Error occurred during mean calculation of a column of a data frame, which is apparently contents numeric data

[Berend Hasselman]

On 29-02-2012, at 09:45, Aniruddha Mukherjee wrote:

> Hello R people,
> 
> How can I compute the mean of the "Pulse_rate" column of the data frame or 
> matrix from the following character object called "str_got". It has 14 
> entries and each entry has 8 values, separated by commas. Please go thru 
> the following R commands to know how I tried to unstring and unlist the 
> values to form a data frame.
>> str_got
> [1] "bp,67,2011-12-09T19:59:44.044+05:30,9830576102,68.0,124.0,58.0,66.0" 
>     "bp,67,2011-12-09T20:19:31.031+05:30,9830576102,72.0,133.0,93.0,40.0" 
> .....
>> 
> matr<-matrix(unlist(strsplit(str_got, ",")), nrows, byrow=T)

nrows? 
I assume this was set somewhere in your script and not shown.
Is it length(str_got)?

>> matr
>        [,1]   [,2]                                              [,3]      
>       [,4]               [,5]        [,6]       [,7]       [,8] 
> [1,] "bp" "67"    "2011-12-09T19:59:44.044+05:30" "9830576102" "68.0" 
> ......

> Note column names must be inserted before computing the desired mean 
> value.
> matr1<-as.data.frame(matr)

Use matr1 <- as.data.frame(matr, stringsAsFactors=FALSE)

If you don't dos tringsAsFactors=FALSE the column will be a factor and that is not equivalent with numeric.

What's wrong with

matr1$Pulse_rate <- as.numeric(matr1$Pulse_rate)

Then you can calculate the desired mean with 

mean(matr1$Pulse_rate)

or

mean(matr1[,"Pulse_rate"])

Berend