syntax-cheatsheet.md

R syntax cheatsheet

Disclaimer: this is just a general outline of things I've found useful.

1. Data structures
2. Numbers
3. Strings
4. General
5. Numerical summaries
6. Plotting
7. Modelling data
8. Probability distributions
9. Useful functions

1. DATA STRUCTURES

Data frame

data.frame(a, b, ...)

Vector

vector() # empty or c(a, b, ...) # specific values or c(a:b) # integer sequence

Matrix

matrix(data, nrow = y, ncol = x)

List

list(a, b, ...)

Display structure

str(data)

... or length

length(data)

Convert data structure

as.vector(data) /matrix, etc.

Rename table column

names(data)[index] <- name

Add row

rbind(data, c(a, b, ...))

... or column

cbind(data, c(a, b, ...))

Append to specific location

append(data, c(a, b, ...), i) # where i is the position before append

Note on indexing

Starts from 1 as opposed to 0;
variable[row, column] (e.g. list first ten rows of matrix data[1:10, ] or first ten columns data[ , 1:10])

2. NUMBERS

Sequence

seq(min, max, increment)

Random numbers (uniform distribution)

runif(n, min, max) # where n is the quantity of values to be generated

Rounding

round(value, dp)

Transpose

t(data)

Cumulative sum

cumsum(data)

3. STRINGS

Concatenate vectors to character strings

paste(x, y, ..., sep = " ", collapse = NULL)

Convert characters from lower to uppercase and vice versa

toupper(data)
tolower(data)
casefold(x, upper = TRUE/FALSE)  # alternative method
chartr(old = "example", new = "eXaMpLe", data)  # specify new character pattern

Convert string to expression

parse(text = data)

4. GENERAL

Define variable

x <- vector() # or x <- a; y <- b; z <- c for a series of expressions

Loops

for (i in seq(min, max)) {
  do something
}

while (condition) {
  do something
}

repeat {
  do something
  if (condition) {
    break
  }
}

Conditionals

if (condition) {
  do a
} else if (condition) {
  do b
} else {
  do c
}

Functions

function(arguments) {
  return(something)
}

(Note that curly brackets are not needed for single-expression functions.)

Manipulate function body

body(f) <- as.call(c(as.name("{"), e)) # where f is an arbitrary function

Manipulate function arguments

formals(f) <- alist(a = , b = , ...) # where f is an arbitrary function; a and b new arguments

Operators

Listed here

Errors

to do: message warning, stop

Return n number of first or last values

head(data, n) or tail(data, n)

Look for object

exists("variable")

Check for duplicates in data

duplicated(data)

# Returns a vector of Boolean values.
# If input is a vector, compares consecutive values;
# and if input is a matrix, compares consecutive row vectors.

Remove duplicates

unique(data)

Read clipboard data (MacOS)

read.table(pipe("pbpaste"), sep = "\t", header = F)

Read line from terminal

readline(prompt = "Enter something: ") # disable quoting in strings by including quote = ""

Substitute values (e.g., decimal commas to dots)

scan(text = x, dec = ",", sep = ".") or as.numeric(sub(",", ".", sub(".", "", var, fixed = TRUE), fixed = TRUE))

Check if argument value was specified to a function

missing(data)

e.g.,

a <- function(x, y) {
  if (missing(y)) {
    do something
  }
}

5. NUMERICAL SUMMARIES

Summary

summary(data)

Measures of location

Mean

mean(data)

Weighted mean

weighted.mean(data, weights) # where the two input variables are arbitrary vectors

(Note that if data contains NAs include argument na.rm = TRUE.)

Median

median(data)

Mode

mode <- function(x) {
  ux <- unique(x)
  ux[which.max(tabulate(match(x, ux)))]
}

mode(data)

Extreme upper

max(data)

Extreme lower

min(data)

Measures of spread

Range

range(data)

IQR

IQR(data)

Variance

var(data)

Standard deviation

sqrt(var(data))

6. PLOTTING

Saving images of plots (in current directory)

png("filename.png")
x()  # where x() is an arbitrary function for producing graphical representations
dev.off()

Stemplot

stem(data, scale = 1)

Simple bar chart

barplot(data, xlab = "", ylab = "", main = "", names.arg = c(), col = "")

Side-by-side bar chart

... add argument beside = TRUE;
modify width of bars with width = 0-1;
and input data may be a table or individual vectors.

Add legend

legend("position", title = "", legend = sort(unique(data$groups)), fill = c(), box.lty = 1, box.lwd = 1, box.col = "")
# where position is defined as coordinates x, y, or right, left, etc.;
# data$groups represents a discrete random variable that describes each group;
# lty refers to type, lwd to width, and col colour.

Frequency histogram

hist(data, xlim = c(), right = FALSE, breaks = c(), main = "", xlab = "", col = "")

(Note that right = FALSE transforms bins from right-closed left-open (a, b] to left-closed right-open [a, b), e.g., will include 5 in bin 5-10 instead of 0-5.)

Unit-area histogram

... add argument freq = FALSE

Boxplot

boxplot(vector, data_frame, horizontal = TRUE, main = "", col = "")

(Note that horizontal = TRUE results in horizontal boxplot.)

Comparative boxplot

boxplot(values ~ groups, data,
  xlab = "",
  ylab = "",
  ylim = c(),
  main = "",
  names = c(),
  horizontal = TRUE,
  notch = FALSE,  # confidence interval around median
  varwidth = TRUE,  # box width (/height in horizontal) proportionate to sample size
  col = c()
)

Scatterplot (to fit a line see below)

plot(x, y,
  xlab = "",
  ylab = "",
  xlim = c(),
  ylim = c(),
  main = ""
)

7. MODELLING DATA

Linear regression

model <- lm(dependent variable ~ independent variable, data = data.frame)

plot(dependent variable ~ independent variable, data = data.frame)

abline(model)

Normal probability plot

qqnorm(data)
qqline(data)

8. PROBABILITY DISTRIBUTIONS

to do

d - for density - density/mass function
p - for probability - cumulative distribution function
q - for quantile - inverse c.d.f.
r - for random - sampling function

Distributions:

Binomial - binom
Geometric - geom
Poisson - pois
Uniform - unif
Normal - norm

9. USEFUL FUNCTIONS

Check out scripts/ folder in the repository.