为什么我在 colMeans(x, na.rm = TRUE) 中出现错误：'x' 必须是数字

Question

My code is as follows:

my_filtered_data <- my_data[, colSums(my_data != 0) >= 300]

set.seed(123)
data1.csv <- my_filtered_data[sample(nrow(my_filtered_data), 200), ]
data2.csv <- data.frame(data1.csv)
data3.csv <- scale(data2.csv, center = TRUE) # Gives error.

Error in colMeans(x, na.rm = TRUE) : 'x' must be numeric

Can someone explain why I am receiving this error?

Answer 1

This is a bit longer for a comment, may or may not answer the problem. But I think this can be one of the issues with OP dataset.

You used data.frame command , now data.frame contains a parameter stringsAsFactors = TRUE by default,and probably this is converting one of your columns to factor, that is why you are getting this error, One way to avoid it to use options(stringsAsFactors=FALSE) on top of your code or use data.frame(your_object, stringsAsFactors=FALSE)

Just to recreate the error , you can use iris dataset to display a similar error,

scale(iris[,1:5], center=TRUE, scale=TRUE) 
## This fails with the same error as the last column in iris data set is a factor

but this will work,

scale(iris[,1:4], center=TRUE, scale=TRUE)

Note I am dropping here the column, In your case you might want to change it to numeric(so it totally depends on what you are trying to do here). In case you do want to change it to numeric from a factor, try running as.numeric(as.character(your_column)) .

Also as suggested in comments , try avoiding names which contains dot in in your object in R.

So, sum of all the info can be wrapped in one line that , you should try this:

data.frame(data1.csv, stringsAsFactors=FALSE) then try running the scale command.