# Composing functions 

DRY (Don't repeat yourself!). If writing essentially the same code more than once, make a function.

Functions should be designed to be simple, to always return the same outputs when given the same
inputs, and to avoid the use of side effects (having effects in addition to returning an output)

## Bad

``` 
#t_output_IT.R
data <- read.data(...)
filter <- "IT"
if (filter == "IT"){
   filtered_data <- data %>% filter(...)
}else if(filter == "SE"){
  filtered_data <- data %>% filter(...)
}
run_analysis(filtered_data)

#t_output_SE.R
data <- read.data(...)
filter <- "SE"
if (filter == "IT"){
   filtered_data <- data %>% filter(...)
}else if(filter == "SE"){
  filtered_data <- data %>% filter(...)
}
run_analysis(filtered_data)
```
## Good

```
#t_output.R
data <- read.data(...)
filter_data <- function(datain, filter){
  if (filter == "IT"){
     return(datain %>% filter(...))
  }
  if(filter == "SE"){
    return(datain %>% filter(...))
  }
}
t_output <- function(datain, filter){
  filtered_data <- filter_data(datain, filter)
  run_analysis(filtered_data)
}
t_output(data, "SE")
t_output(data, "IT")
```

Note that in this example I could probably take my filter_data function and use it in multiple
different files now, while before I would need to copy and paste it everywhere!

# Code style

Truncate all lines at 80 characters (In RStudio: Tools - Global Options - Code - Display - Show Margin (Margin column = 80))

Make blank lines between code chunks (eg, between functions, etc) for better readability

Use comments to say WHY not WHAT.

## Naming functions/objects

Try not to name things like df1, df2, temp1, temp2, etc. Take time to name things more properly (not easy!)

Suggestion: for functions, use verb_noun construction, e.g. derive_baseline, get_data

Avoid using "." in R names, as S3 methods make use of the ".", so this can confuse.