---
title: "Working with labelled data"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Working with labelled data}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
```

```{r}
library(rio)
library(haven)
```

Formats SAS, SPSS, and Stata  use `haven` as import and export functions. And these formats can have the so-called labelled data. For more information, please read `vignette("semantics", "haven")`. Here, we provide a quick guide on how to work with labelled data using `rio`.

You can use `haven::labelled()` to create labelled data.

```{r}
gender <- haven::labelled(
                     c("M", "F", "F", "F", "M"),
                     c(Male = "M", Female = "F"))
```

Or directly using `attrs`

```{r}
rating <- sample(1:5)
attr(rating, "labels") <-  c(c(Good = 1, Bad = 5))
```

```{r}
mydata <- data.frame(gender, rating)
```

Round trip: The data labels are retained. But they are at the variable level.

```{r}
export(mydata, "mydata.sav")
restored_data <- rio::import("mydata.sav")
str(restored_data)
```

`rio::gather_attrs()` converts attributes to the data.frame level

```{r}
g <- rio::gather_attrs(restored_data)
str(g)
attr(g, "labels")
```

```{r include = FALSE}
unlink("mydata.sav")
```