Question

I have two large vectors:

A: https://dl.dropbox.com/u/22681355/A.csv
B: https://dl.dropbox.com/u/22681355/B.csv

A has over 20000 entries but only 1350 unique entries.
B is a random number generated from 1 to 9 exactly 1350 times

I would like to assign values from B to A such that the same values in A get the same values in B. e.g. if there are multiple 1’s each 1 should get the same number from B.

I have been using the A[B] command but after the 18000th entry I get NAs

What is the proper way of doing this?

code:

A<-read.csv("A.csv")
B<-read.csv("B.csv")

A[B]

Answer 1

1. read.csv() creates a data frame, not a vector.
2. You probably mean B[A] which for each element in A gets the value of B at the index of that element’s value. Since A’s values range from 1 to 1899 it exceeds B’s size of 1349. For those elements outside the bounds of B, NAs get introduced.

The correct way to doing what you want to achieve is

A = read.table("http://dl.dropbox.com/u/22681355/A.csv")
B = read.table("http://dl.dropbox.com/u/22681355/B.csv")
A = A$V1
B = B$V1
A = as.factor(A)

B[match(A,levels(A))]

match(A,levels(A)) will return a vector of the same length as A that for each element contains the position of the element of A in its factor’s levels, i.e. a number between 1 and 1350 (1350 distinct values). If A was as.factor(c(1,1,3,5,5,7)), levels(A) would be c(1,3,5,7) and match(A,levels(A)) would be c(1,1,2,3,3,4), i.e. the position of the element in it’s levels.