Iteration over 2 lists in R-CodePudding

I would like to do an iteration with 2 lists.

For a single case, I have one dataframe df1 and one vector v1.

My reproducible example as below.

df1 <- data.frame(n1 = c(2,2,0),
                 n2 = c(2,1,1),
                 n3 = c(0,1,1),
                 n4 = c(0,1,1))

v1 <- c(1,2,3)

Now, I calculate an value (ses.value) for each row using this code

x <- (v1 - apply(df1, 1, mean))/apply(df1,1,sd)

Let's say we will have a list of multiple dataframes l1 and a list of vectors l2 (each list has the same number of elements) Now, I would like to run a loop for those lists by using the above code (the element of l1 must go with the element of l2 with the same position).

# 3 dataframes and 3 vectors
df1 <- data.frame(n1 = c(2,2,0), n2 = c(2,1,1), n3 = c(0,1,1), n4 = c(0,1,1))
df2 <- data.frame(n1 = c(1,6,0), n2 = c(2,1,8), n3 = c(0,2,1), n4 = c(0,7,1))
df3 <- data.frame(n1 = c(1,6,0), n2 = c(9,1,5), n3 = c(4,2,1), n4 = c(0,7,2))

v1 <- c(1,2,3)
v2 <- c(2,3,4)
v3 <- c(4,5,6)

# list
l1 <- list(df1,df2,df3)
l2 <- list(v1,v2,v3)

Since my lists are too big, using for loop might be not such a good idea, any suggestions using lapply or something similar?

CodePudding user response：

Since your lists apparently are large, you probably could benefit from rowMeans2 and rowSds of the matrixStats package.

library(matrixStats)
Map(\(x, y) (y - rowMeans2(as.matrix(x))) / rowSds(as.matrix(x)), l1, l2)
# [[1]]
# [1] 0.0 1.5 4.5
# 
# [[2]]
# [1]  1.3055824 -0.3396831  0.4057513
# 
# [[3]]
# [1] 0.1237179 0.3396831 1.8516402

Data:

l1 <- list(structure(list(n1 = c(2, 2, 0), n2 = c(2, 1, 1), n3 = c(0, 
1, 1), n4 = c(0, 1, 1)), class = "data.frame", row.names = c(NA, 
-3L)), structure(list(n1 = c(1, 6, 0), n2 = c(2, 1, 8), n3 = c(0, 
2, 1), n4 = c(0, 7, 1)), class = "data.frame", row.names = c(NA, 
-3L)), structure(list(n1 = c(1, 6, 0), n2 = c(9, 1, 5), n3 = c(4, 
2, 1), n4 = c(0, 7, 2)), class = "data.frame", row.names = c(NA, 
-3L)))

l2 <- list(c(1, 2, 3), c(2, 3, 4), c(4, 5, 6))

CodePudding user response：

We can use Map to loop over the corresponding elements of each list and then do the calculation based on OP's code

Map(function(x, y) (y - apply(x, 1, mean))/apply(x,1,sd), l1, l2)

-output

[[1]]
[1] 0.0 1.5 4.5

[[2]]
[1]  1.3055824 -0.3396831  0.4057513

[[3]]
[1] 0.1237179 0.3396831 1.8516402

Also, if the datasets are really big, use dapply from collapse, which is more efficient

library(collapse)
 Map(function(x, y) (y - dapply(x, MARGIN = 1,
     FUN = fmean))/dapply(x, MARGIN = 1, FUN = fsd), l1, l2)