How to read and write STATA variable and value labels from R.

snippets.R

# Using read.dta() to load a sample STATA data file
dta <- read.dta("my_toy_data.dta")

# Let's describe dta
str(dta)
#'data.frame':  95202 obs. of  13 variables:
#    $ hh     : num  1.01e+09 1.01e+09 1.01e+09 1.01e+09 1.01e+09 ...
#$ h14aq2 : Factor w/ 61 levels "matooke","matooke",..: 50 51 34 47 57 41 22 1 13 53 ...
#$ h14aq3 : Factor w/ 93 levels "Kilogram (kg)",..: 51 26 85 1 NA 1 55 70 1 33 ...
#$ h14aq4 : num  0.5 4 4 1 NA 1.5 5 3 4 10 ...
#$ h14aq5 : num  125 2000 800 1600 1000 1200 1500 3000 2800 1000 ...
#$ h14aq6 : num  NA NA NA NA NA NA NA NA NA NA ...
#$ h14aq7 : num  NA NA NA NA NA NA NA NA NA NA ...
#$ h14aq8 : num  NA NA NA NA NA NA NA NA NA NA ...
#$ h14aq9 : num  NA NA NA NA NA NA NA NA NA NA ...
#$ h14aq10: num  NA NA NA NA NA NA 4 NA NA NA ...
#$ h14aq11: num  NA NA NA NA NA NA 1200 NA NA NA ...
#$ h14aq12: num  250 500 200 1600 NA 800 300 1000 700 100 ...
#$ h14aq13: num  NA NA NA NA NA NA NA NA NA NA ...
#- attr(*, "datalabel")= chr ""
#- attr(*, "time.stamp")= chr ""
#- attr(*, "formats")= chr  "%11.0f" "%8.0f" "%8.0f" "%8.2f" ...
#- attr(*, "types")= int  102 102 102 102 102 102 102 102 102 102 ...
#- attr(*, "val.labels")= chr  "" "h14aq2" "h14aq3" "" ...
#- attr(*, "var.labels")= chr  "household code" "code" "unit of quantity" "hh consumption out of purchase.quantity" ...
#- attr(*, "version")= int 7
#- attr(*, "label.table")=List of 2
#..$ h14aq2: Named num  101 102 103 104 105 106 107 108 109 110 ...
#.. ..- attr(*, "names")= chr  "matooke" "matooke" "matooke" "matooke" ...
#..$ h14aq3: Named num  1 2 3 4 5 6 7 8 9 10 ...
#.. ..- attr(*, "names")= chr  "Kilogram (kg)" "Gram" "Litre" "Small cup with handle (Akendo)" ...


# "val.labels", "var.labels", "label.table" are attributes specific to the STATA .dta format
# In this example they contain:
attr(dta, "var.labels")
#[1] "household code"                                     
#[2] "code"                                               
#[3] "unit of quantity"                                   
#[4] "hh consumption out of purchase.quantity"            
#[5] "hh consumption out of purchase. value"              
#[6] "consumption out of purchase away from home.quantity"
#[7] "consumption out of purchase away from home.value"   
#[8] "consumption out of home produce quantity"           
#[9] "consumption out of home produce value"              
#[10] "received in kind.quantity"                          
#[11] "received in kind.value"                             
#[12] "market price"                                       
#[13] "farm gate price"  

attr(dta, "val.labels")
#[1] ""       "h14aq2" "h14aq3" ""       ""       ""       ""       ""      
#[9] ""       ""       ""       ""       ""  
# So there are value labels attached to columns 2:3 ("code" and "unit of quantity")

attr(dta, "label.table")$h14aq2
#matooke                            matooke 
#101                                102 
#matooke                            matooke 
#103                                104 
#sweet potatoes fresh               sweet potatoes dry 
#105                                106 
#cassava fresh                      cassava dry/flour 
#107                                108 
# etc...

attr(dta, "label.table")$h14aq3
#Kilogram (kg)                      Gram 
#1                                  2 
#Litre                              Small cup with handle (Akendo) 
#3                                  4 
#Metre                              Square metre 
#5                                  6 
#Yard                               Millilitre 
#7                                  8 
# etc...

# Attach variable and value labels to a new R data.frame
attr(df, "var.labels") <- c("my var label 1", "my var label 2")

# To create and attach value labels
val1 <- c(apple=10, beans=20)
val2 <- c(kg=1, bag=2)
attr(df, "label.table") <- list(val1, val2)
attr(df, "val.labels") <- c("val1", "val2")

# And finally describe the dataset
attr(df, "datalabel") <- "Collected from MWI 2013"

# Author and date are nice to have too
attr(df, "time.stamp") <- Sys.Date()
attr(df, "author") <- "Melanie Bacou, mel@mbacou.com"

# When using write.dta(df), STATA will read in the attached datalabel, time.stamp, and value labels.
# Depending on STATA version it might also read in the variable labels (I'm having mixed luck here).

Hi,

thank you for your post.

I am trying to implement something very similar to this, but I am having troubles. Specifically, I realized that the function write.dta has been deprecated and the most immediate replacement (save.dta13) does not work. Also, in line 85 of your script, it seems that a vector of strings is being attached to the dataset attributes. Is that the correct procedure?

I actually tried to apply your script to my case, but when I read the result in Stata, it does not show value labels.

Many thanks,
NF

@florenzion I came across your post as I am having trouble exporting value labels from R->STATA as well. Did you ever find a solution to this?