error in right_join when both by variables have label attributes

[rkingdc]

I'm getting the error:

Error: cannot join on columns 'i' x 'i': Can't join on 'i' x 'i' because of incompatible types (numeric / numeric)

When the by variable i from both datasets have a label attribute pulled in from haven::read_sas.

Reproducible example via attached .sas7bdat files:

options(stringsAsFactors=FALSE)
library(haven)
library(dplyr)

tbl_right <- read_sas('right.sas7bdat')
tbl_left  <- read_sas('left.sas7bdat')

right_join(tbl_right, tbl_left, by='i')

attr(tbl_right$i, 'label') <- NULL
attr(tbl_left$i, 'label') <- NULL

right_join(tbl_right, tbl_left, by='i')

> sessionInfo()
R version 3.1.2 (2014-10-31)
Platform: x86_64-w64-mingw32/x64 (64-bit)

locale:
[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252   
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] dplyr_0.4.3 haven_0.2.0

loaded via a namespace (and not attached):
[1] assertthat_0.1 DBI_0.3.1      magrittr_1.5   parallel_3.1.2 R6_2.1.0       Rcpp_0.11.6   
[7] tools_3.1.2

example_files.zip

[romainfrancois]

Can you make this reproducible. We don't have your .sas7bdat files

[rkingdc]

I attached a .zip file. Are they not in there?

[romainfrancois]

Ok thanks. I did not see the attachment before.

[romainfrancois]

Ok so what happens is that we don't know how to join these so it fails. We could instead drop attributes, but maybe this would lead to other issues. ping @hadley ?

[huftis]

When the attributes are identical (which they typically are if the two data frames come from the same source), this could be handled gracefully, by just including the attributes on the resulting data frame.

If they are not identical, keep the first set of attributes (i.e. the ones from the source data frame) and issue a warning that the second ones have been dropped?

[hadley]

@ikkyle can you please make an inline reproducible example? You shouldn't need to use haven or external data

[rkingdc]

library(dplyr)

# I had to make the "by" variables numeric rather than integer--it works fine when they're integers
tbl_left <- tbl_df(data.frame(i = rep(c(1.0, 2.0, 3.0), each = 2),
                              x1 = letters[1:6]))
tbl_right <- tbl_df(data.frame(i = c(1.0, 2.0, 3.0),
                               x2 = letters[1:3]))

left_join(tbl_left, tbl_right, by = 'i')

attr(tbl_left$i, 'label') <- 'iterator'
attr(tbl_right$i, 'label') <- 'iterator'

left_join(tbl_left, tbl_right, by = 'i')

[hadley]

• For future reference, please don't include session info unless explicitly asked
• Can you please use data_frame() instead of tbl_df() + data.frame()

[rkingdc]

tbl_left <- data_frame(
  i = rep(c(1.0, 2.0, 3.0), each = 2),
  x1 = letters[1:6]
)
tbl_right <- data_frame(
  i = c(1.0, 2.0, 3.0),
  x2 = letters[1:3]
)

left_join(tbl_left, tbl_right, by = 'i')

attr(tbl_left$i, 'label') <- 'iterator'
attr(tbl_right$i, 'label') <- 'iterator'

left_join(tbl_left, tbl_right, by = 'i')

[hadley]

@romainfrancois this at least needs a better error message.

I think @huftis has the right strategy if the attributes are identical, but if they're not identical, I think we might need more than a warning.