De STATA a R: uso de las funciones del paquete labelled

labelled.R

``` r
library(tidyverse)
library(haven)
library(labelled)

# crear data frame ------------------------------

set.seed(22)

dt_nolab <- tibble(age = floor(rnorm(6,25,10)), 
                   sex = gl(2,1,6, labels = c("f","m")),
                   edu = rep(c("primaria","secundaria"),each=3)) %>% 
  
  #transformar vectores chr a fct
  mutate_if(is.character,as.factor) 

# labels? -----------------------------------------------------------------

Hmisc::contents(dt_nolab)
#> 
#> Data frame:dt_nolab  6 observations and 3 variables    Maximum # NAs:0
#> 
#> 
#>     Levels Storage
#> age         double
#> sex      2 integer
#> edu      2 integer
#> 
#> +--------+-------------------+
#> |Variable|Levels             |
#> +--------+-------------------+
#> |   sex  |f,m                |
#> +--------+-------------------+
#> |   edu  |primaria,secundaria|
#> +--------+-------------------+
var_label(dt_nolab)
#> $age
#> NULL
#> 
#> $sex
#> NULL
#> 
#> $edu
#> NULL
look_for(dt_nolab)
#>   variable
#> 1      age
#> 2      sex
#> 3      edu

#asignar etiquetas a variables ------------------
dt <- dt_nolab %>% 
  labelled::set_variable_labels(
    age = "Edad (años)", 
    sex = "Sexo de participantes",
    edu = "Nivel educativo") 

# verificar labels --------------------------------------------------------

Hmisc::contents(dt)
#> 
#> Data frame:dt    6 observations and 3 variables    Maximum # NAs:0
#> 
#> 
#>                    Labels Levels Storage
#> age           Edad (años)         double
#> sex Sexo de participantes      2 integer
#> edu       Nivel educativo      2 integer
#> 
#> +--------+-------------------+
#> |Variable|Levels             |
#> +--------+-------------------+
#> |   sex  |f,m                |
#> +--------+-------------------+
#> |   edu  |primaria,secundaria|
#> +--------+-------------------+
var_label(dt)
#> $age
#> [1] "Edad (años)"
#> 
#> $sex
#> [1] "Sexo de participantes"
#> 
#> $edu
#> [1] "Nivel educativo"
look_for(dt_nolab)
#>   variable
#> 1      age
#> 2      sex
#> 3      edu
look_for(dt_nolab,details = T)
#>   variable   class    type               levels value_labels unique_values
#> 1      age numeric  double                                               6
#> 2      sex  factor integer                 f; m                          2
#> 3      edu  factor integer primaria; secundaria                          2
#>   n_na na_values na_range
#> 1    0                   
#> 2    0                   
#> 3    0

str(dt)
#> Classes 'tbl_df', 'tbl' and 'data.frame':    6 obs. of  3 variables:
#>  $ age: num  19 49 35 27 22 43
#>   ..- attr(*, "label")= chr "Edad (años)"
#>  $ sex: Factor w/ 2 levels "f","m": 1 2 1 2 1 2
#>   ..- attr(*, "label")= chr "Sexo de participantes"
#>  $ edu: Factor w/ 2 levels "primaria","secundaria": 1 1 1 2 2 2
#>   ..- attr(*, "label")= chr "Nivel educativo"
attributes(dt$age)
#> $label
#> [1] "Edad (años)"
attributes(dt$age)$label
#> [1] "Edad (años)"

# grabar en dta ------------------------------

#dt %>% haven::write_dta("data/new_db.dta")

# leer dta con labels -----------------------------------------------------

#dt_new <- read_dta("data/new_db.dta") %>% as_factor()
dt_new <- dt

Hmisc::contents(dt_new)
#> 
#> Data frame:dt_new    6 observations and 3 variables    Maximum # NAs:0
#> 
#> 
#>                    Labels Levels Storage
#> age           Edad (años)         double
#> sex Sexo de participantes      2 integer
#> edu       Nivel educativo      2 integer
#> 
#> +--------+-------------------+
#> |Variable|Levels             |
#> +--------+-------------------+
#> |   sex  |f,m                |
#> +--------+-------------------+
#> |   edu  |primaria,secundaria|
#> +--------+-------------------+
var_label(dt_new)
#> $age
#> [1] "Edad (años)"
#> 
#> $sex
#> [1] "Sexo de participantes"
#> 
#> $edu
#> [1] "Nivel educativo"
look_for(dt_new)
#>   variable                 label
#> 1      age           Edad (años)
#> 2      sex Sexo de participantes
#> 3      edu       Nivel educativo

# dictionary --------------------------------------------------------

#caso 1: si tienes un df con labels
#extraerlos con var_label()
dt_dict_list <- var_label(dt_new)

# formato común de diccionario
look_for(dt_new,details = T)
#>   variable                 label   class    type               levels
#> 1      age           Edad (años) numeric  double                     
#> 2      sex Sexo de participantes  factor integer                 f; m
#> 3      edu       Nivel educativo  factor integer primaria; secundaria
#>   value_labels unique_values n_na na_values na_range
#> 1                          6    0                   
#> 2                          2    0                   
#> 3                          2    0

#caso 2: df con variable name y variable label
#transformar a formato lista
dt_dict_wide_list <- 
  look_for(dt_new) %>% 
  as_tibble() %>% 
  pivot_wider(names_from = variable,values_from = label) %>% 
  as.list() %>% 
  print()
#> $age
#> [1] "Edad (años)"
#> 
#> $edu
#> [1] "Nivel educativo"
#> 
#> $sex
#> [1] "Sexo de participantes"

# from data to list  --------------------------------------------------

dt_dict_list
#> $age
#> [1] "Edad (años)"
#> 
#> $sex
#> [1] "Sexo de participantes"
#> 
#> $edu
#> [1] "Nivel educativo"
dt_dict_wide_list
#> $age
#> [1] "Edad (años)"
#> 
#> $edu
#> [1] "Nivel educativo"
#> 
#> $sex
#> [1] "Sexo de participantes"

# assing data labels ------------------------------------------------------

dt_nolab <- dt_nolab %>% 
  mutate(log=c(rep(T,5),F)) %>% 
  select(sex,log,age,edu)

# limitante:
# requieres tener todas las variables de la lista en la base de datos
# al activar select(-sex) se obtiene error

Hmisc::contents(dt_nolab)
#> 
#> Data frame:dt_nolab  6 observations and 4 variables    Maximum # NAs:0
#> 
#> 
#>     Levels Storage
#> sex      2 integer
#> log        logical
#> age         double
#> edu      2 integer
#> 
#> +--------+-------------------+
#> |Variable|Levels             |
#> +--------+-------------------+
#> |   sex  |f,m                |
#> +--------+-------------------+
#> |   edu  |primaria,secundaria|
#> +--------+-------------------+

dt_nolab_newlab <- dt_nolab %>% 
  mutate(log=c(rep(T,5),F)) %>% 
  select(sex,log,age,edu) %>% 
  #select(-sex) %>% #limitante
  #set_variable_labels(.labels = dt_dict_list) #en caso 1
  set_variable_labels(.labels = dt_dict_wide_list) #en caso 2

Hmisc::contents(dt_nolab_newlab)
#> 
#> Data frame:dt_nolab_newlab   6 observations and 4 variables    Maximum # NAs:0
#> 
#> 
#>                    Labels Levels Storage
#> sex Sexo de participantes      2 integer
#> log                              logical
#> age           Edad (años)         double
#> edu       Nivel educativo      2 integer
#> 
#> +--------+-------------------+
#> |Variable|Levels             |
#> +--------+-------------------+
#> |   sex  |f,m                |
#> +--------+-------------------+
#> |   edu  |primaria,secundaria|
#> +--------+-------------------+
look_for(dt_nolab_newlab) %>% as_tibble()
#> # A tibble: 4 x 2
#>   variable label                
#>   <chr>    <chr>                
#> 1 sex      Sexo de participantes
#> 2 log      <NA>                 
#> 3 age      Edad (años)          
#> 4 edu      Nivel educativo
```

<sup>Created on 2019-08-08 by the [reprex package](https://reprex.tidyverse.org) (v0.3.0)</sup>