Remove same columns from left_join

Question

I\'d like to merge two data frames by id, but they both have 2 of the same columns; therefore, when I merge i get new .x and .y column

Answer 1

Just drop everything you don't want from df2 - in this case the id and value2 columns:

left_join(df1, select(df2, c(id,value2)), by = "id")

#  id     value1 element day     value2
#1  1  1.2276303   TEST1  15 -0.1389861
#2  2 -0.8017795   TEST1  15 -0.5973131
#3  3 -1.0803926   TEST1  15 -2.1839668
#4  4 -0.1575344   TEST1  15  0.2408173
#5  5 -1.0717600   TEST1  15 -0.2593554

Beware that not all these answers are equivalent, and ask what it is you need as a result. E.g.:

df1 <- data.frame(id=1:3,day=2:4,element=3:5,value1=100:102)
df2 <- data.frame(id=1:3,day=3:5,element=4:6,value2=200:202)
df1

#  id day element value1
#1  1   2       3    100
#2  2   3       4    101
#3  3   4       5    102

df2
#  id day element value2
#1  1   3       4    200
#2  2   4       5    201
#3  3   5       6    202

left_join(df1, df2)
#Joining by: c("id", "day", "element")
#  id day element value1 value2
#1  1   2       3    100     NA
#2  2   3       4    101     NA
#3  3   4       5    102     NA

left_join(df1, select(df2, c(id,value2)), by = "id")
#  id day element value1 value2
#1  1   2       3    100    200
#2  2   3       4    101    201
#3  3   4       5    102    202

Answer 2

df <- left_join(df1, df2, by = c("id", "element", "day"))

Answer 3

You only need:

df <- left_join(df1, df2)

by = NULL, the default, join will do a natural join, using all variables with common names across the two tables. A message lists the variables so that you can check they're right

Output:

Joining by: c("id", "element", "day")
  id     value1 element day     value2
1  1 -0.6264538   TEST1  15 -0.8204684
2  2  0.1836433   TEST1  15  0.4874291
3  3 -0.8356286   TEST1  15  0.7383247
4  4  1.5952808   TEST1  15  0.5757814
5  5  0.3295078   TEST1  15 -0.3053884

It's worth pointing out the comment by thelatemail: "Joining on id is not the same as joining on id/element/day". However, in this specific example, because element and day are the same for all records in both tables we get the same result.

Original result

Data

set.seed(1)
df1 <- data.frame(id = seq(1,5), value1 = rnorm(5), element = "TEST1", day = 15) 
df2 <- data.frame(id = seq(1,5), value2 = rnorm(5), element = "TEST1", day = 15) 
df <- left_join(df1, df2, by = "id")

Output:

  id     value1 element.x day.x     value2 element.y day.y
1  1 -0.6264538     TEST1    15 -0.8204684     TEST1    15
2  2  0.1836433     TEST1    15  0.4874291     TEST1    15
3  3 -0.8356286     TEST1    15  0.7383247     TEST1    15
4  4  1.5952808     TEST1    15  0.5757814     TEST1    15
5  5  0.3295078     TEST1    15 -0.3053884     TEST1    15

Answer 4

After having checked that these columns are indeed the same, you could just remove them before doing the join

if (all(df1[,c('element', 'day')] == df2[,c('element', 'day')]))
  df <- left_join(df1[,setdiff(colnames(df1),c('element', 'day'))], df2, by = "id")
else
  stop("Should not happen!?")