mis_val(,with_labels = TRUE) incompatible with mutate()

[bhaney22]

The "with_labels = TRUE" option in the "mis_val()" command changes the class of the variable to "labelled". This variable no longer be used as input into mutate() commands.

See example below:

I read in the GSS.sav file

mygss <- as.data.frame(read_sav("GSS2018.sav")) %>%
select(YEAR,ID,WRKSTAT,SEX,RACE,RINCOME,INCOME,WEALTH,REGION,
AGE,MARITAL,CHILDS,EDUC,DEGREE,
PARTYID,PRES16,NATEDUC,COURTS,TAX,NATFARE,NATENRGY,PARTTEAM,LIFE)

Here is the structure of the variable before recording with mis_val()

str(mygss$NATENRGY)
dbl+lbl [1:2348] 1, 2, 1, 1, 1, 2, 2, 1, 2, 1, 2, 2, 2, 1, 2, NA, 1, N...
@ label : chr "Developing alternative energy sources"
@ format.spss : chr "F1.0"
@ display_width: int 10
@ labels : Named num [1:6] 0 1 2 3 8 9
..- attr(*, "names")= chr [1:6] "IAP" "Too little" "About right" "Too much" …

mis_val does not change type if with_labels option is not used

mygss$NATENRGY <- mis_val(mygss$NATENRGY, c(0,8,9))

str(mygss$NATENRGY)
dbl+lbl [1:2348] 1, 2, 1, 1, 1, 2, 2, 1, 2, 1, 2, 2, 2, 1, 2, NA, 1, N...
@ label : chr "Developing alternative energy sources"
@ format.spss : chr "F1.0"
@ display_width: int 10
@ labels : Named num [1:6] 0 1 2 3 8 9
..- attr(*, "names")= chr [1:6] "IAP" "Too little" "About right" "Too much" …

The recoded variable works fine with mutate:

mygss <- mygss %>%

• mutate(renew = case_when(NATENRGY == 1 ~ 1,

•                       NATENRGY != 1 ~ 0)) 
  

calc_cro(mygss,NATENRGY,renew)

| | | renew | |

		0	1
Developing alternative energy sources	IAP
	Too little		1277
	About right	788
	Too much	169
	DON'T KNOW
	No answer
	#Total cases	957	1277

But, the labels are left in there, so I use the with_labels = TRUE option

mis_val using the with_labels = TRUE option changes the class of the variable

mygss$NATENRGY <- mis_val(mygss$NATENRGY, c(0,8,9),with_labels = TRUE)

Now the variable is of the 'labelled' class

str(mygss$NATENRGY)
Class 'labelled' dbl+lbl [1:2348] 1, 2, 1, 1, 1, 2, 2, 1, 2, 1, 2, 2, 2, 1, 2, NA, 1, N...
@ format.spss : chr "F1.0"
@ display_width: int 10
.. .. LABEL: Developing alternative energy sources
.. .. VALUE LABELS [1:3]: 1=Too little, 2=About right, 3=Too much

labelled class variables cannot be used in the mutate() commands

mygss <- mygss %>%

• mutate(renew = case_when(NATENRGY == 1 ~ 1,

•                       NATENRGY != 1 ~ 0)) 
  

Error: Problem with mutate() input renew.
x Can't combine ..1 <labelled> and ..2 .
ℹ Input renew is case_when(NATENRGY == 1 ~ 1, NATENRGY != 1 ~ 0).

[gdemin]

Hi,
Thank you for report.
I can't reproduce your problem with latest dplyr. Could you provide result of the sessionInfo()?

[bhaney22]

Thank you for your quick response! Here is the result of sessionInfo() just after running the error causing command:

sessionInfo()
R version 3.6.3 (2020-02-29)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.4 LTS

Matrix products: default
BLAS: /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.7.1
LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.7.1

locale:
[1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C LC_TIME=en_US.UTF-8
[4] LC_COLLATE=en_US.UTF-8 LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
[7] LC_PAPER=en_US.UTF-8 LC_NAME=C LC_ADDRESS=C
[10] LC_TELEPHONE=C LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C

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

other attached packages:
[1] plm_2.2-3 dynlm_0.3-6 visualize_4.3.0 effects_4.1-3
[5] AER_1.2-7 survival_3.2-3 sandwich_2.5-1 lmtest_0.9-37
[9] zoo_1.8-8 car_3.0-8 carData_3.0-4 expss_0.10.5
[13] ggformula_0.9.4.9002 ggridges_0.5.2 scales_1.1.1 ggstance_0.3.4
[17] ggplot2_3.3.2 stargazer_5.2.2 DataExplorer_0.8.1 wooldridge_1.3.1
[21] dplyr_1.0.0 haven_2.3.1 readr_1.3.1

loaded via a namespace (and not attached):
[1] nlme_3.1-148 matrixStats_0.56.0 tools_3.6.3 backports_1.1.8
[5] R6_2.4.1 DBI_1.1.0 lazyeval_0.2.2 colorspace_1.4-1
[9] nnet_7.3-14 withr_2.2.0 tidyselect_1.1.0 gridExtra_2.3
[13] curl_4.3 compiler_3.6.3 cli_2.0.2 htmlTable_2.0.1
[17] mosaicCore_0.6.0 checkmate_2.0.0 stringr_1.4.0 digest_0.6.25
[21] foreign_0.8-76 minqa_1.2.4 rmarkdown_2.3 rio_0.5.16
[25] pkgconfig_2.0.3 htmltools_0.5.0 bibtex_0.4.2.2 lme4_1.1-23
[29] htmlwidgets_1.5.1 rlang_0.4.7 readxl_1.3.1 rstudioapi_0.11
[33] farver_2.0.3 generics_0.0.2 zip_2.0.4 magrittr_1.5
[37] Formula_1.2-3 Matrix_1.2-18 fansi_0.4.1 Rcpp_1.0.5
[41] munsell_0.5.0 abind_1.4-5 lifecycle_0.2.0 stringi_1.4.6
[45] gbRd_0.4-11 MASS_7.3-51.6 plyr_1.8.6 grid_3.6.3
[49] parallel_3.6.3 bdsmatrix_1.3-3 forcats_0.5.0 crayon_1.3.4
[53] lattice_0.20-41 splines_3.6.3 hms_0.5.3 knitr_1.29
[57] pillar_1.4.6 igraph_1.2.5 boot_1.3-25 glue_1.4.1
[61] packrat_0.5.0-25 evaluate_0.14 mitools_2.4 data.table_1.12.8
[65] vctrs_0.3.2 nloptr_1.2.2.2 tweenr_1.0.1 Rdpack_0.11-0
[69] miscTools_0.6-26 networkD3_0.4 cellranger_1.1.0 gtable_0.3.0
[73] purrr_0.3.4 polyclip_1.10-0 tidyr_1.1.0 assertthat_0.2.1
[77] xfun_0.15 ggforce_0.3.2 openxlsx_4.1.5 survey_3.36
[81] tibble_3.0.3 maxLik_1.3-8 statmod_1.4.34 ellipsis_0.3.1

[bhaney22]

UPDATE:

Here is a minimum working example to illustrate the error:
`
library(dplyr)
library(haven)
library(expss)

var1 <- haven::labelled(c(0, 1, 9, 0, 0),c(No = 0, Yes = 1, DK = 9))
var2 <- haven::labelled(c(0, 1, 2, 3, 1), c(DK = 0, first = 1, second = 2, third = 3))

mytest <- tibble(var1,var2)

mytest$var1 <- expss::mis_val(mytest$var1, c(9))
mytest$var2 <- expss::mis_val(mytest$var2, c(0), with_labels = TRUE)

This mutate command works fine

mytest <- mytest %>%
mutate(newvar1 = case_when(var1 == 1 ~ 1,
var1 != 1 ~ 0))

This mutate command causes the error

mytest <- mytest %>%
mutate(newvar2 = case_when(var2 == 1 ~ 1,
var2 != 1 ~ 0))
`

Here is the error I get:

Error: Problem with mutate() input newvar2.
x Can't combine ..1 <labelled> and ..2 .
ℹ Input newvar2 is case_when(var2 == 1 ~ 1, var2 != 1 ~ 0).

Here is the sessionInfo() results:

sessionInfo()
R version 3.6.3 (2020-02-29) Platform: x86_64-pc-linux-gnu (64-bit) Running under: Ubuntu 18.04.4 LTS Matrix products: default BLAS: /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.7.1 LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.7.1 locale: [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C LC_TIME=en_US.UTF-8 [4] LC_COLLATE=en_US.UTF-8 LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8 [7] LC_PAPER=en_US.UTF-8 LC_NAME=C LC_ADDRESS=C [10] LC_TELEPHONE=C LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C attached base packages: [1] stats graphics grDevices utils datasets methods base other attached packages: [1] haven_2.3.1 expss_0.10.5 dplyr_1.0.0 loaded via a namespace (and not attached): [1] Rcpp_1.0.5 rstudioapi_0.11 knitr_1.29 magrittr_1.5 [5] hms_0.5.3 tidyselect_1.1.0 R6_2.4.1 rlang_0.4.7 [9] fansi_0.4.1 stringr_1.4.0 tools_3.6.3 packrat_0.5.0-25 [13] htmlTable_2.0.1 checkmate_2.0.0 data.table_1.12.8 xfun_0.15 [17] utf8_1.1.4 cli_2.0.2 matrixStats_0.56.0 htmltools_0.5.0 [21] ellipsis_0.3.1 assertthat_0.2.1 digest_0.6.25 tibble_3.0.3 [25] lifecycle_0.2.0 crayon_1.3.4 readr_1.3.1 purrr_0.3.4 [29] htmlwidgets_1.5.1 vctrs_0.3.2 glue_1.4.1 stringi_1.4.6 [33] compiler_3.6.3 pillar_1.4.6 forcats_0.5.0 generics_0.0.2 [37] backports_1.1.8 foreign_0.8-76 pkgconfig_2.0.3
--
 
  |  
 

[bhaney22]

Since this is a combination of packages causing the error, I also submitted the following request to dplyr:

tidyverse/dplyr#5424 (comment)

[gdemin]

Thanks, now I can reproduce the issue. Currently I don't understand why it is happen. It is not about labelled or haven_labelled class. It seems it is something connected with vctrs_vctr class inside dplyr. There is no error when I remove this class:

library(dplyr)
library(haven)
library(expss)

var2 <- haven::labelled(c(0, 1, 2, 3, 1), c(DK = 0, first = 1, second = 2, third = 3))

mytest <- tibble(var2)

mytest$var2 <- expss::mis_val(mytest$var2, c(0), with_labels = TRUE)

class(mytest$var2)
# remove vctrs_vctr class
class(mytest$var2) = setdiff(class(mytest$var2), "vctrs_vctr")

# works as expected
mytest <- mytest %>%
    mutate(newvar2 = case_when(var2 == 1 ~ 1,
                               var2 != 1 ~ 0))

I will try to make a workaround for this issue in the next version.
As quick fix you can use read_spss from expss:

library(dplyr)
library(haven)
library(expss)

mygss <- expss::read_spss("~/GSS2018.sav") %>%
    select(YEAR,ID,WRKSTAT,SEX,RACE,RINCOME,INCOME,WEALTH,REGION,
           AGE,MARITAL,CHILDS,EDUC,DEGREE,
           PARTYID,PRES16,NATEDUC,COURTS,TAX,NATFARE,NATENRGY,PARTTEAM,LIFE)


mygss$NATENRGY <- mis_val(mygss$NATENRGY, c(0,8,9),with_labels = TRUE)

mygss = mutate(mygss, renew = case_when(NATENRGY == 1 ~ 1,
                                        NATENRGY != 1 ~ 0)) 

calc_cro(mygss,NATENRGY,renew)

or, if you want to go with haven:

library(dplyr)
library(haven)
library(expss)

mygss <- haven::read_spss("~/GSS2018.sav") %>%
    select(YEAR,ID,WRKSTAT,SEX,RACE,RINCOME,INCOME,WEALTH,REGION,
           AGE,MARITAL,CHILDS,EDUC,DEGREE,
           PARTYID,PRES16,NATEDUC,COURTS,TAX,NATFARE,NATENRGY,PARTTEAM,LIFE)

mygss = add_labelled_class(mygss, remove_classes = c("haven_labelled", "vctrs_vctr" ))

mygss$NATENRGY <- mis_val(mygss$NATENRGY, c(0,8,9),with_labels = TRUE)

mygss = mutate(mygss, renew = case_when(NATENRGY == 1 ~ 1,
                                        NATENRGY != 1 ~ 0)) 

calc_cro(mygss,NATENRGY,renew)

[bhaney22]

Thank you for your quick response and helpful feedback! I agree that this seems like an error caused by different approaches to labels by haven in the tidyverse and in expss. As I mentioned in a previous comment, I did submit this issue to dplyr. I try to work within the tidyverse as much as possible, which is why I use haven to read in the SPSS file. But, there is no tidyverse package for creating the wide variety of human-readable tables that expss provides. Thank you again for providing the expss package.