`bind_rows` fails to bind data frames with haven labels

[cimentadaj]

I'm trying to bind three data frames which have only one column with haven labels. The values of the columns are characters and the actual labels are numbers. Below is a reprex:

library(haven)
library(dplyr)
labs1 <- c(`92` = "British Virgin Islands", `184` = "Cook Islands")
labs2 <- c(`20` = "Andorra", `660` = "Anguilla")
labs3 <- c(`16` = "American Samoa", `212` = "Dominica")

vec1 <- labelled(rep("Cook Islands", 5), labs1)
vec2 <- labelled(rep("Andorra", 5), labs2)
vec3 <- labelled(rep("Dominica", 5), labs3)

df1 <- data.frame(LocName = vec1)
df2 <- data.frame(LocName = vec2)
df3 <- data.frame(LocName = vec3)
df_all <- list(df1, df2, df3)

bind_rows(df_all)
#> Error: Can't convert from `$LocName` <labelled<character>> to `$LocName` <labelled<character>> due to loss of precision.
#> * Locations: 1, 2, 3, 4, 5
#> Values are labelled in `$LocName` but not in `$LocName`.
do.call(rbind, df_all)
#> Error: Can't convert from <labelled<character>> to <labelled<character>> due to loss of precision.
#> * Locations: 1, 2, 3, 4, 5
#> Values are labelled in `` but not in ``.

The expected output would be one single data frame with the labels merged into one final column. I'm using latest dplyr and haven on CRAN.

SI: ```r devtools::session_info() #> ─ Session info ─────────────────────────────────────────────────────────────── #> setting value #> version R version 4.0.2 (2020-06-22) #> os Ubuntu 20.04.1 LTS #> system x86_64, linux-gnu #> ui X11 #> language (EN) #> collate en_US.UTF-8 #> ctype en_US.UTF-8 #> tz Europe/Berlin #> date 2020-08-24 #> #> ─ Packages ─────────────────────────────────────────────────────────────────── #> package * version date lib source #> assertthat 0.2.1 2019-03-21 [1] CRAN (R 4.0.0) #> backports 1.1.8 2020-06-17 [1] CRAN (R 4.0.2) #> callr 3.4.3 2020-03-28 [1] CRAN (R 4.0.0) #> cli 2.0.2 2020-02-28 [1] CRAN (R 4.0.0) #> crayon 1.3.4 2017-09-16 [1] CRAN (R 4.0.0) #> desc 1.2.0 2018-05-01 [1] CRAN (R 4.0.0) #> devtools 2.3.0 2020-04-10 [1] CRAN (R 4.0.2) #> digest 0.6.25 2020-02-23 [1] CRAN (R 4.0.0) #> dplyr 1.0.2 2020-08-18 [1] CRAN (R 4.0.2) #> ellipsis 0.3.1 2020-05-15 [1] CRAN (R 4.0.0) #> evaluate 0.14 2019-05-28 [1] CRAN (R 4.0.0) #> fansi 0.4.1 2020-01-08 [1] CRAN (R 4.0.0) #> forcats 0.5.0 2020-03-01 [1] CRAN (R 4.0.0) #> fs 1.5.0 2020-07-31 [1] CRAN (R 4.0.2) #> generics 0.0.2 2018-11-29 [1] CRAN (R 4.0.0) #> glue 1.4.1 2020-05-13 [1] CRAN (R 4.0.0) #> haven 2.3.1 2020-06-01 [1] CRAN (R 4.0.2) #> highr 0.8 2019-03-20 [1] CRAN (R 4.0.0) #> hms 0.5.3 2020-01-08 [1] CRAN (R 4.0.0) #> htmltools 0.5.0 2020-06-16 [1] CRAN (R 4.0.1) #> knitr 1.29 2020-06-23 [1] CRAN (R 4.0.2) #> lifecycle 0.2.0 2020-03-06 [1] CRAN (R 4.0.0) #> magrittr 1.5 2014-11-22 [1] CRAN (R 4.0.0) #> memoise 1.1.0 2017-04-21 [1] CRAN (R 4.0.0) #> pillar 1.4.6 2020-07-10 [1] CRAN (R 4.0.2) #> pkgbuild 1.1.0 2020-07-13 [1] CRAN (R 4.0.2) #> pkgconfig 2.0.3 2019-09-22 [1] CRAN (R 4.0.0) #> pkgload 1.1.0 2020-05-29 [1] CRAN (R 4.0.0) #> prettyunits 1.1.1 2020-01-24 [1] CRAN (R 4.0.0) #> processx 3.4.3 2020-07-05 [1] CRAN (R 4.0.2) #> ps 1.3.4 2020-08-11 [1] CRAN (R 4.0.2) #> purrr 0.3.4 2020-04-17 [1] CRAN (R 4.0.0) #> R6 2.4.1 2019-11-12 [1] CRAN (R 4.0.0) #> Rcpp 1.0.5 2020-07-06 [1] CRAN (R 4.0.2) #> remotes 2.1.1 2020-02-15 [1] CRAN (R 4.0.0) #> rlang 0.4.7 2020-07-09 [1] CRAN (R 4.0.2) #> rmarkdown 2.2 2020-05-31 [1] CRAN (R 4.0.0) #> rprojroot 1.3-2 2018-01-03 [1] CRAN (R 4.0.0) #> sessioninfo 1.1.1 2018-11-05 [1] CRAN (R 4.0.0) #> stringi 1.4.6 2020-02-17 [1] CRAN (R 4.0.0) #> stringr 1.4.0 2019-02-10 [1] CRAN (R 4.0.0) #> testthat 2.3.2 2020-03-02 [1] CRAN (R 4.0.0) #> tibble 3.0.3.9000 2020-07-29 [1] Github (tidyverse/tibble@b4eec19) #> tidyselect 1.1.0 2020-05-11 [1] CRAN (R 4.0.0) #> usethis 1.6.1 2020-04-29 [1] CRAN (R 4.0.0) #> vctrs 0.3.2 2020-07-15 [1] CRAN (R 4.0.2) #> withr 2.2.0 2020-04-20 [1] CRAN (R 4.0.0) #> xfun 0.16 2020-07-24 [1] CRAN (R 4.0.2) #> yaml 2.2.1 2020-02-01 [1] CRAN (R 4.0.0) #> #> [1] /usr/local/lib/R/site-library #> [2] /usr/lib/R/site-library #> [3] /usr/lib/R/library ```

[cportner]

I am running into a similar issue where the labeled variables created from Stata files by haven produces an error when trying to bind rows. Is there a preferred way to deal with this until the underlying problem is fixed? zap_labels and as_factor will both do the trick. It does not look like haven offers an option to ignore labels when using read_dta.

library(haven)
library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
labs1 <- c(`92` = "British Virgin Islands", `184` = "Cook Islands")
labs2 <- c(`20` = "Andorra", `660` = "Anguilla")
labs3 <- c(`16` = "American Samoa", `212` = "Dominica")

vec1 <- labelled(rep("Cook Islands", 5), labs1)
vec2 <- labelled(rep("Andorra", 5), labs2)
vec3 <- labelled(rep("Dominica", 5), labs3)

df1 <- data.frame(LocName = vec1)
df2 <- data.frame(LocName = vec2)
df3 <- data.frame(LocName = vec3)
df_all <- list(df1, df2, df3)

bind_rows(zap_labels(df_all))
#> Error: Can't convert from `$LocName` <labelled<character>> to `$LocName` <labelled<character>> due to loss of precision.
#> * Locations: 1, 2, 3, 4, 5
#> Values are labelled in `$LocName` but not in `$LocName`.

zap_df1 <- zap_labels(df1)
zap_df2 <- zap_labels(df2)
zap_df3 <- zap_labels(df3)

bind_rows(zap_df1, zap_df2, zap_df3)
#>         LocName
#> 1  Cook Islands
#> 2  Cook Islands
#> 3  Cook Islands
#> 4  Cook Islands
#> 5  Cook Islands
#> 6       Andorra
#> 7       Andorra
#> 8       Andorra
#> 9       Andorra
#> 10      Andorra
#> 11     Dominica
#> 12     Dominica
#> 13     Dominica
#> 14     Dominica
#> 15     Dominica

fac_df1 <- as_factor(df1)
fac_df2 <- as_factor(df2)
fac_df3 <- as_factor(df3)

bind_rows(fac_df1, fac_df2, fac_df3)
#>    LocName
#> 1      184
#> 2      184
#> 3      184
#> 4      184
#> 5      184
#> 6       20
#> 7       20
#> 8       20
#> 9       20
#> 10      20
#> 11     212
#> 12     212
#> 13     212
#> 14     212
#> 15     212

^{Created on 2020-10-14 by the reprex package (v0.3.0)}

[hadley]

@gorcha @mikmart any chance either of you could take a look at this?

[gorcha]

Hey @hadley, it looks like this is working correctly at the moment - when figuring out attributes, vec_cast():

• Takes the labels from the to vector if they exist, but the x vector otherwise - note that it doesn't combine them if there are label sets on both.
• Complains about a lossy cast if any values in the x vector are labelled in x but not in to, since we're losing information.

Here's a simpler reprex:

library(haven)
library(vctrs)

vec1 <- labelled(1, c("Good" = 1))
vec2 <- labelled(5, c("Bad" = 5))
vec_cast(vec1, vec2)
#> Error: Can't convert from <labelled<double>> to <labelled<double>> due to loss of precision.
#> * Locations: 1
#> Values are labelled in `` but not in ``.

Automatically combining labels across two different vectors is a bit fraught.
It would be possible, albeit a bit fiddly, but there are some odd edge case issues that could come up that make me question whether it makes sense conceptually (e.g. is it right to add labels for a value that exists in both vectors but is only labelled in one, since there's no guarantee the labels apply in both cases?).

I'm assuming this is why @mikmart chose to drop labels altogether and return the underlying type where they don't match in vec_ptype2().

It also doesn't intuitively feel right to keep labels from x when we're casting to a different object. vec_cast() is implicitly saying "x should become like to", so it doesn't really make sense to combine the labels from x.

So basically, I think it makes more sense to leave it to the user to manually combine/set the labels where this is the case, since there are too many potential edge cases to deal with and at least a few scenarios I can think of where the correct choice depends on user intent.

Maybe what it needs is a coercion vignette that shows what happens when coercing, combining or casting labelled vectors in various situations? Happy to volunteer to put something together.

[hadley]

What if we just combined the labels when they didn't overlap? It seems like that should be safe?

[hadley]

Ah no, there is something going wrong here:

library(haven)
x1 <- labelled("Cook Islands", c(`92` = "British Virgin Islands", `184` = "Cook Islands"))
x2 <- labelled("Andorra", c(`20` = "Andorra", `660` = "Anguilla"))
x3 <- labelled("Dominica", c(`16` = "American Samoa", `212` = "Dominica"))

vctrs::vec_c(x1, x2)
#> [1] "Cook Islands" "Andorra"

vctrs::vec_c(x1, x2, x3)
#> Error: Can't convert from <labelled<character>> to <labelled<character>> due to loss of precision.
#> * Locations: 1
#> Values are labelled in `` but not in ``.
vctrs::vec_c(vctrs::vec_c(x1, x2), x3)
#> <labelled<character>[3]>
#> [1] Cook Islands Andorra      Dominica    
#> 
#> Labels:
#>           value label
#>  American Samoa    16
#>        Dominica   212
vctrs::vec_c(x1, vctrs::vec_c(x2, x3))
#> <labelled<character>[3]>
#> [1] Cook Islands Andorra      Dominica    
#> 
#> Labels:
#>                   value label
#>  British Virgin Islands    92
#>            Cook Islands   184

^{Created on 2021-04-09 by the reprex package (v2.0.0)}