JMM JMM - 2 months ago 13
R Question

Correlation between two corresponding columns from seperate datasets

I have two sets of data, which contain columns with the same names, but differing values in those columns. e.g:

m1 <- matrix(1:9, nrow = 3, ncol = 3, byrow = TRUE,
dimnames = list(c("s1", "s2", "s3"),c("cow", "dog","cat")))
m2 <- matrix(1:9, nrow = 3, ncol = 3, byrow = FALSE,
dimnames = list(c("s1", "s2", "s3"),c("dog", "cow","cat")))
> m1
cow dog cat
s1 1 2 3
s2 4 5 6
s3 7 8 9
> m2
dog cow cat
s1 1 4 7
s2 2 5 8
s3 3 6 9


I would like to create a function using cor.test() to calculate the correlation between corresponding columns. E.g. cow vs cow, dog vs dog. The reason for using cor.test() is I want to obtain the correlation coefficient and p-value. So, if there are other ways to obtain this information, i'm open to those too. The actual data set has thousands of columns, which are randomly organized, so im looking for a way to match the columns first and then calculate the correlation. Any ideas?

Answer

Here is a solution, using lapply on common columns:

# Common columns
cols <- intersect(colnames(m1), colnames(m2))

# For each column, compute cor test
res <- lapply(cols, function(x) cor.test(
  m1[, x],
  m2[, x]
))

names(res) <- cols

The result is a list of htest objects that you can access this way: res[["cow"]]