Merging two data frames with unique keys

Question

I want to select rows of data1 if the value of the first column exists in data2:

data1<-data.table(a=c(1,2,3,1),b=c(1,4,6,3))
data2<-data.table(a=c(1,3,1,5,6))


Desired output:
a b 
1 1 
3 6 
1 3

The merge doesn't work as the keys are not unique, any other idea?

+3

Soheil May 17 '15 at 10:54

2 answers

We can create a sequence column ("ind") in each dataset for column "a" and "join" after setting the key column as "a" and "ind"

data1[, ind:=1:.N, a]
data2[, ind:= 1:.N, a]
setkey(data1, a, ind)[data2, nomatch=0][, ind:=NULL]
#   a b
#1: 1 1
#2: 3 6
#3: 1 3

+4

akrun May 17 '15 at 10:58

nrussell · Accepted Answer · 2015-05-17T10:59:58+0000

Here's a no joins approach:

> unique(data1[a %in% data2[,a],])
   a b
1: 1 1
2: 3 6
3: 1 3