R Analytics: "K-means clustering" option of "R statistics operation" block broken?

joshlk · May 7, 2015 12:21PM

When executing the "K-means clustering" option of "R statistics operation" block an error is raised:

"Error executing R operation.

Error in data.frame(input_data, cluster) :
arguments imply differing number of rows: 139, 61
Execution halted"

I had a look at the underlying R code and it looks like the data is first filtered for null values:

"subset<-na.omit(subset)"<br />
But the last operation tries to remerge the original dataset with the clustering results even though some of the original lines where omitted.

Versions -
Omniscope: 2.9 build 1593 x64
R version: 3.2.0

glen · May 8, 2015 10:33AM

Hi Josh,

You are quite right about the source of the error. The problem is on the last line of the script (line 22):

output_data <- data.frame(input_data,cluster)

If you replace line 22 with this code your script will work:

output_data<-data.frame(input_data[!apply((sapply(input_data[fields],is.na)),1,any),], cluster)

We will also amend the error in the script for an upcoming build!

Sincerely,

Glen

Welcome!

Categories

Ideas Parade

Tagged