Reputation: 902
I have a matrix with NA and I want to find the mean of every 3 columns for each row.
set.seed(100)
a <- matrix(sample(c(NA, 1:4), 90, replace = TRUE), 10)
If I use
t(apply(a, 1, tapply, gl(3, 3), mean))
The code cannot find the average for colums including NA. I wonder how I can modify the code. Thanks a lot.
Upvotes: 1
Views: 579
Reputation: 887241
We can do this by also converting to a 3D array and then get the rowMeans
after looping through the 3rd MARGIN.
res <- apply(array(a, c(nrow(a), 3, 3)), 3, rowMeans, na.rm = TRUE)
all.equal(res, t(apply(a, 1, tapply, gl(3, 3),
mean, na.rm = TRUE)), check.attributes = FALSE)
#[1] TRUE
Upvotes: 1