L12_Factors.Rmd

---
title: "Factors"
author: "Bert Gollnick"
output: 
  html_document:
    toc: true
    toc_float: true
    code_folding: hide
    number_sections: true
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = T, warning = T, message = T)
```

# Introduction

A factor is very similar to a vector. The difference is that only a fixed list of values (so called levels) is allowed.

# Creation

Factors store categorical values.

```{r}
x <- factor(c("a", "b", "c", "b"))
levels(x)
```

# Changing Values and Levels

You can reassign values.

```{r}
x[1] <- "c"
```

But you cannot use values that are not defined in the levels.

```{r}
x[1] <- "d"
x
```

You can't combine factors

```{r}
factor1 <- factor("first")
factor2 <- factor("second")
c(factor1, factor2)
```

When should you use it? --> When you know the allowed range of values:

```{r}
worktype_char <- c("employed", "unemployed", "employed")
# if you want to keep status "retired" as well, you need to add it as a level, even if your initial data does not have it.
worktype_factor <- factor(x = worktype_char, 
                          levels = c("employed", "unemployed", "retired"))
```

This way you can get information on that level as well:

```{r}
table(worktype_factor)
```

# Data Import Problem

When you import data and the column has some non-numeric values, the column type will be set to factor. One reason might be a wrong decimal separator.

```{r}
txt_data <- read.table("./data/factorImport.txt", 
                       header = T)
txt_data
```

# Casting Factors to Numeric

If you want to cast it to numeric, you need to cast it to character at first!

Otherwise you get the factor levels instead of the values.

```{r}
# how NOT to do it
levels(txt_data$value)
as.numeric(txt_data$value) 
```

```{r}
# how to do it
as.numeric(as.character(txt_data$value))
```