useful_r_commands.Rmd

---
title: "Useful R Commands"
author: "Martin Oberg"
date: "2023-03-29"
output: html_document
---

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

This is a collection of R commands that do useful things

# Reading Data

## Load Multiple JSON Files
```{r}
library(purrr)
library(tidyverse)
library(jsonlite)

path <- "./your_path"
files <- dir(path, pattern = "*.json")

# ~fn(.) allows files to be %>%'d into the '.' position
data <- files %>% map_df(~fromJSON(file.path(path, .), flatten = TRUE))
# map_df supersceeded? Use purrr::list_rbind instead?
```



# Operations Across Rows

## Row Differences

Calculate differences between values.
```{r}
df = tibble(ts=c(10, 20, 30, 40))
# dplyr::lag returns the lagged value, which is subtracted from ts to get the difference
df %>% dplyr::mutate(d = ts - dplyr::lag(ts, n=1))
```


## Time Differences Across Rows

Here we have timestamped (ts) data with an event trigger and response.
We want to calculate the time between a trigger and the nearest following target.
```{r}
library(tidyverse)
df = tibble(ts=c(2,4,6,8,10,12,14), trigger=c(0,1,1,0,1,0,0), response=c(0,0,1,0,0,1,1))
# make groups of triggered events by performing a cumulative sum on trigger
df <- df %>% mutate(idx=cumsum(trigger))
df
```


```{r}
# Clean up
df <- df %>% 
    # make groups of triggered events by performing a cumulative sum on trigger
    mutate(idx=cumsum(trigger)) %>% 
    # number responses for each trigger event
    group_by(idx) %>% 
    mutate(response=cumsum(response)*response) %>% 
    ungroup() %>% 
    # remove unneeded rows
    filter(trigger == 1 | response == 1)

df %>% 
    group_by(idx) %>% 
    summarize(
        idx = unique(idx),
        trigger_ts = ts[match(1, trigger)],
        response_ts = ts[match(1, response)]
    )
```

# Conditions Across Columns

## Does a Column Contain a value?

```{r}
df <- tibble(col1=c('zza', 'zyz', 'zzz'), col2=c('xxa', 'xax', 'xyz'))

# map2_lgl uses col1 and "a" as arguments for str_detect
df %>%
    mutate(col1_has_a = map2_lgl(col1, "a", str_detect))

# if_any operates over columns (might need to select columns to be in a particular
# order if using a subset, or use a tidy_select() command)
df %>%
    mutate(col1or2_has_a = if_any(col1:col2, ~ str_detect(., "a")))
```