Join columns mislabelled

[ShixiangWang]

Thanks for making this nice tool firstly. I got an error when I tried to rewrite code with dtplyr. I debugged the code and found the left_join() returns the result of right join and vice versa.

library(dplyr, warn.conflicts = FALSE)
library(data.table)
library(dtplyr)


a <- lazy_dt(data.table(x_left = 1:3, y = 1))
b <- lazy_dt(data.table(x_right = 2:4, z = 2))

left_join(a, b, by=c("x_left"="x_right"))

Join result

> left_join(a, b, by=c("x_left"="x_right"))
Source: local data table [?? x 4]
Call:   `_DT2`[`_DT1`, on = .(x_right = x_left), allow.cartesian = TRUE]

  x_right     z     y
    <int> <dbl> <dbl>
1       1    NA     1
2       2     2     1
3       3     2     1

# Use as.data.table()/as.data.frame()/as_tibble() to access results

> left_join(as_tibble(a), as_tibble(b), by=c("x_left"="x_right"))
# A tibble: 3 x 3
  x_left     y     z
   <int> <dbl> <dbl>
1      1     1    NA
2      2     1     2
3      3     1     2

[hadley]

Can you please provide a minimal reproducible example using the reprex package? The goal of a reprex is to make it as easy as possible for me to recreate your problem so that I can fix it. If you've never made a minimal reprex before, there is lots of good advice here.

[ShixiangWang]

@hadley Sorry for not using reprex before because I thought it is easy to reproduce.

library(dplyr, warn.conflicts = FALSE)
library(data.table)
#> 
#> 载入程辑包：'data.table'
#> The following objects are masked from 'package:dplyr':
#> 
#>     between, first, last
library(dtplyr)


a <- lazy_dt(data.table(x_left = 1:3, y = 1))
b <- lazy_dt(data.table(x_right = 2:4, z = 2))

left_join(a, b, by=c("x_left"="x_right"))
#> Source: local data table [?? x 4]
#> Call:   `_DT2`[`_DT1`, on = .(x_right = x_left), allow.cartesian = TRUE]
#> 
#>   x_right     z     y
#>     <int> <dbl> <dbl>
#> 1       1    NA     1
#> 2       2     2     1
#> 3       3     2     1
#> 
#> # Use as.data.table()/as.data.frame()/as_tibble() to access results
left_join(as_tibble(a), as_tibble(b), by=c("x_left"="x_right"))
#> # A tibble: 3 x 3
#>   x_left     y     z
#>    <int> <dbl> <dbl>
#> 1      1     1    NA
#> 2      2     1     2
#> 3      3     1     2

^{Created on 2019-12-09 by the reprex package (v0.3.0)}

[hadley]

Minimal reprex:

library(data.table)
library(dtplyr)
library(dplyr, warn.conflicts = FALSE)


x1 <- lazy_dt(data.table(x_left = 1:3, y = 1))
x2 <- lazy_dt(data.table(x_right = 2:4, z = 2))

left_join(x1, x2, by = c("x_left" = "x_right"))
#> Source: local data table [?? x 4]
#> Call:   `_DT2`[`_DT1`, on = .(x_right = x_left), allow.cartesian = TRUE]
#> 
#>   x_right     z     y
#>     <int> <dbl> <dbl>
#> 1       1    NA     1
#> 2       2     2     1
#> 3       3     2     1
#> 
#> # Use as.data.table()/as.data.frame()/as_tibble() to access results

^{Created on 2019-12-24 by the reprex package (v0.3.0)}

I don't think that you're getting a right join (because if that was the case you'd expect x_right to have values 2:4, but instead the column name is being taken from the right hand side instead of the left hand side.

[ShixiangWang]

@hadley Thanks for correcting my question and fixing this bug :)