Sample.Split in R SplitRatio parameter has to be i [0,1]

Question

I am getting the following error:

Error in sample.split: 'SplitRatio' parameter has to be i [0, 1] range or [1, length(Y)] range

when I try to run the following code:

set.seed(1000)
library(caTools)
split = sample.split(letters$isB, SplitRatio = 0.5)

How do I get this function to work? My syntax does not appear to be incorrect. — Superad, Feb 23 '16 at 00:22
Beats me. It'd be a lot easier to diagnose if it were reproducible, but `letters$isB` is not valid. — Adam Hoelscher, Feb 23 '16 at 01:01
Make sure you've dropped or omitted all NA values so that you have same vector length. This will also cause this error. — Antonio, Aug 23 '22 at 16:54

score 10 · Answer 1 · answered Aug 29 '16 at 15:43

10

There is nothing wrong with the syntax. You probably spelled your outcome variable (letters$isB) incorrectly. Since letters$isB does not exist (or not loaded), you get that error.

answered Aug 29 '16 at 15:43

Mahesh Mitikiri

126
2
5

score 1 · Answer 2 · edited Feb 28 '18 at 20:04

I received an error similar to what was listed above. I realized that I had forgotten to change my variable in the code listed below

split = sample.split(dataset$Profit,

from profit to units sold (variable in my actual data set) vs. profit which was code from another project. Hope this helps - I listed the rest of my code and my errors below.

> library(caTools)
> set.seed(123)
> split = sample.split(dataset$Profit, SplitRatio = .8)
Error in sample.split(dataset$Profit, SplitRatio = 0.8) : 
  Error in sample.split: 'SplitRatio' parameter has to be i [0, 1] range or [1, length(Y)] range
> training_set = subset(dataset, split == TRUE)
Error in split == TRUE : 
  comparison (1) is possible only for atomic and list types
> test_set = subset(dataset, split == FALSE)
Error in split == FALSE : 
  comparison (1) is possible only for atomic and list types

score 0 · Answer 3 · edited Mar 02 '16 at 15:06

0

Maybe letters$isB length is 0?

edited Mar 02 '16 at 15:06

Alessandro Cuttin

3,822
1
30
36

answered Feb 29 '16 at 18:05

Juanje

1

score 0 · Answer 4 · edited May 23 '17 at 12:34

The previous answer (https://stackoverflow.com/a/35706404/6188234)

"maybe letters$isB length is 0?"

makes sense with more context. In my experience with another MOOC I received this error, and came to SO looking for an answer.

After referring to Sample.split in R - SplitRatio parameter

I noted that the error is generated if the variable you are trying to split doesn't exist - because of a typo. So the error message misleads the coder to look at the SplitRatio Constant, instead of the variable you are splitting.

split = sample.split(letters$THISDOESNOTEXIST, SplitRatio = 0.5)

In my case this typo was the camelCase of the variable name, so it was difficult to seethe syntax error. Fixing that type cleared this error.

I hope this works for you.

score 0 · Answer 5 · answered Feb 05 '18 at 18:05

I have exactly the same problem, and I am sure nothing is wrong with the syntax nor with the variables. More interestingly, the code works if I run the related chunk manually on Rmarkdown, but when I run the whole markdown from top to bottom, it returns error.

score 0 · Answer 6 · answered Sep 21 '18 at 08:29

0

set.seed(1000) library(caTools) split = sample.split(letters$isB, SplitRatio = 0.5)

isB should be the label of the Dependent variable, look up in your dataset that name.

Here you can find why this error is raised.

answered Sep 21 '18 at 08:29

Anant

396
4
11

score 0 · Answer 7 · answered Aug 23 '22 at 16:53

0

I had this same problem. Make sure you've dropped or omitted all your NA values.

answered Aug 23 '22 at 16:53

Antonio

417
2
8

Sample.Split in R SplitRatio parameter has to be i [0,1]

7 Answers7