"Error in validate_dta_label(label)" when using haven write

I'm having an issue with the write_dta function from the package haven.

My code goes like this:

lab_attr <- label$lab 


attr(data, "label") <- lab_attr
attributes(data)

write_dta(data,"donnees/test_haven_1_data.dta", label = attr(data, "label"))

Which gives me the following error:

Error in validate_dta_label(label) : length(label) == 1 is not TRUE

All of my labels are under 80 characters.

Does anyone have any idea on how to fix this?

Thank you!

Have a nice day,

Cassandra

CodePudding user response：

The way to store Stata variable labels as a data frame attribute is to assign to the attributes of each column, rather than of the data frame.

For example, using the iris data:

data  <- iris

names(data)
# "Sepal.Length" "Sepal.Width"  "Petal.Length" "Petal.Width"  "Species"

If you just wanted to edit one label you could do:

attr(data$Sepal.Length, "label")  <- "new label"

If you want to create labels that explain what the variables are in a loop, you can do:

# Set labels which replace each . in the column name with a space
for(col in names(data)) {
    attr(data[[col]], "label")  <- gsub("\\.", " ", col)
}
# Print attributes
sapply(data, attr, "label")

#   Sepal.Length    Sepal.Width   Petal.Length    Petal.Width        Species 
# "Sepal Length"  "Sepal Width" "Petal Length"  "Petal Width"      "Species"

Once this is done, you do not need to set the label argument in write_dta(). It automatically writes the label attribute to the Stata labels. So in your case you can just do:

write_dta(data,"donnees/test_haven_1_data.dta")