Agenda

1. Vectors and Matrices
2. Lists and Data Frames
3. Type Coercion
4. Troubleshooting

Data Types in R

• R has 5 main atomic data types:
  • Logical
  • Integer
  • Numeric
  • Complex
  • Character

Data Structures in R

1. One-dimensional:
   • Vectors
   • Lists
2. Multi-dimensional:
   • Matrices
   • Data frames

• Everything in R is an object
• Objects can have attributes
  • e.g., names, dimension, length

Vectors in R

• A vector is the most basic structure in R
• It is one-dimensional; its single dimension is its length
• A vector of length n has n cells
• Each cell can hold a single value, like a numeric or string value
  • In general, vectors can only hold ONE type of data

numVec <- c(2,3,4)      # <- is the assigning operator
numVec

## [1] 2 3 4

Examples of Vectors

Examples of character, logical, and complex vectors:

intVec <- c(2L, 3L, 4L)
intVec

## [1] 2 3 4

charVec <- c("red", "green", "blue")
charVec

## [1] "red"   "green" "blue"

logVec <- c(TRUE, FALSE, FALSE, T, F)
logVec

## [1]  TRUE FALSE FALSE  TRUE FALSE

Matrices

• A matrix is a special case of a vector
  • Unlike vectors, matrices have a dimension attribute

myMat <- matrix(nrow = 2, ncol = 4)
myMat

##      [,1] [,2] [,3] [,4]
## [1,]   NA   NA   NA   NA
## [2,]   NA   NA   NA   NA

attributes(myMat)

## $dim
## [1] 2 4

Matrices

myMat <- matrix(1:8, nrow = 2, ncol = 4)
myMat # matrices are filled in column-wise

##      [,1] [,2] [,3] [,4]
## [1,]    1    3    5    7
## [2,]    2    4    6    8

A matrix is a special case of a vector

myVec <- 1:8
myVec

## [1] 1 2 3 4 5 6 7 8

dim(myVec) <- c(2,4)
myVec

##      [,1] [,2] [,3] [,4]
## [1,]    1    3    5    7
## [2,]    2    4    6    8

• Similar to vectors, all elements of a matrix should be of the same data type
  • If not, R automatically coerces

Other Ways to Create a Matrix

• Intuitively, matrices seem to be a combination of vectors that are put next to each other (either column-wise or row-wise).

• rbind() and cbind() (row bind and column bind) achieve this:

vec1 <- 1:4
vec2 <- sample(1:100, 4, replace = FALSE)
vec3 <- sample(1:20, 4, replace=TRUE)
colMat <- cbind(vec1, vec2, vec3)
colMat

##      vec1 vec2 vec3
## [1,]    1   50   15
## [2,]    2   72   18
## [3,]    3   15    8
## [4,]    4   93   19

Other Ways to Create a Matrix

vec1 <- 1:4
vec2 <- sample(1:100, 4, replace = TRUE)
vec3 <- sample(1:20, 4, replace=FALSE)
rowMat <- rbind(vec1, vec2, vec3)
rowMat

##      [,1] [,2] [,3] [,4]
## vec1    1    2    3    4
## vec2   86   94   82   60
## vec3   14   16    3    7

Factors

• A factor is a vector used to specify a discrete classification (categorical values).
• Factors can be ordered or un-ordered
• Levels of a factor are better when labeled (self-descriptive)
  • Consider sex as (0, 1) as opposed to labeled (“F”, “M”)

Sex <- rep(c("Female", "Male"), times = 3)
Sex

## [1] "Female" "Male"   "Female" "Male"   "Female" "Male"

SexFac1 <- factor(Sex)
SexFac1

## [1] Female Male   Female Male   Female Male  
## Levels: Female Male

Factors

levels(SexFac1)

## [1] "Female" "Male"

table(SexFac1)

## SexFac1
## Female   Male 
##      3      3

Factors

SexFac1 # levels are ordered alphabetically - 1st level = BaseLevel

## [1] Female Male   Female Male   Female Male  
## Levels: Female Male

SexFac2 <- factor(Sex, levels = c("Male", "Female"))
SexFac1

## [1] Female Male   Female Male   Female Male  
## Levels: Female Male

SexFac2

## [1] Female Male   Female Male   Female Male  
## Levels: Male Female

Lists

• Think of a list as a vector with the following main differences:
  • Each element of a list can have its own data structure regardless of other elements
    • vector, matrix, another list
  • This means, each element can be of a different data type and a different length

myVec <- c(10, "R", 5L, T)
myVec

## [1] "10"   "R"    "5"    "TRUE"

Lists

myList <- list(10, "R", 5L, T)
myList

## [[1]]
## [1] 10
## 
## [[2]]
## [1] "R"
## 
## [[3]]
## [1] 5
## 
## [[4]]
## [1] TRUE

• Elements of a list are shown with [[]]
• Elements of a vector are shown with []

Data Frames

• A data frame is a special list where all objects have equal length
• A data frame looks very similar to a matrix; however, different columns in a data frame can be different data types

studentID <- paste("S#", sample(c(6473:7392), 10), sep = "")
score <- sample(c(0:100), 10)
sex <- sample(c("female", "male"), 10, replace = TRUE)
data <- data.frame(studentID = studentID, score = score, sex = sex)
head(data)

##   studentID score    sex
## 1    S#7019    40 female
## 2    S#6968     9   male
## 3    S#7025    14   male
## 4    S#6972    73 female
## 5    S#7320    78   male
## 6    S#7279    79 female

Special Values

There are some special values in R:

• Use L to refer to an integer value, e.g., 1L
• R knows infinity: Inf, -Inf
• NaN: refers to “Not a number”

intVec <- c(1L, 2L, 3L, 4L) 
intVec

## [1] 1 2 3 4

typeof(intVec)

## [1] "integer"

intVec*Inf

## [1] Inf Inf Inf Inf

a <- Inf; b <- 0
rslt <- c(b/a, a/a)
rslt

## [1]   0 NaN

Missing Values

• There are two kinds of missing values in R:
  • NaN: stands for “Not a Number” and is a missing value produced by numerical computation.
  • NA: stands for “Not Available” and is used when a value is missing
• NaN is also considered as NA (the reverse is NOT true)

a <- c(1,2)
a[3]

## [1] NA

b <- 0/0
b

## [1] NaN

Data Type Coercion

• In general, vectors CANNOT have mixed types of objects

numCharVec <- c(3.14, "a")
numCharVec                 # What do you expect to be printed? 

numLogVec <- c(pi, T)
numLogVec                   

charLogVec <- c("a", TRUE)
charLogVec 

• The above are examples of implicit coercion
• Explicit coercion is also possible

Data Type Coercion

• as(): explicitly coerces objects from one type to another

numVec <- seq(from = 1200, to = 1300, by = 15)
numVec

## [1] 1200 1215 1230 1245 1260 1275 1290

numToChar <- as(numVec, "character")
numToChar

## [1] "1200" "1215" "1230" "1245" "1260" "1275" "1290"

numToChar==as.character(numVec)

## [1] TRUE TRUE TRUE TRUE TRUE TRUE TRUE

logVec <- c(F, T, F, T, T)
as(logVec, "numeric")

## [1] 0 1 0 1 1

as.numeric(logVec)

## [1] 0 1 0 1 1

Data Type Coercion

• Coercion does not always work! Be careful about warnings:

charVec <- c("2.5", "3", "2.8", "1.5", "zero")
as(charVec, "numeric")

## Warning in asMethod(object): NAs introduced by coercion

## [1] 2.5 3.0 2.8 1.5  NA

charVec <- c("2.5", "3", "2.8", "1.5", zero)

## Error in eval(expr, envir, enclos): object 'zero' not found

Troubleshooting

• Troubleshooting is a vital skill for anyone using a programming language
• Assume things will always break.

1. Try to replicate the error. If you know what types of input will cause an error and which types won’t, this is a clue.
2. Narrow down on where the error is occuring. This typically involves running chunks of code line-by-line or block-by-block.
3. Try fixing the error.
4. Google
   • copy and paste your error/warning message
   • CRAN (Comprehensive R Archive Network)
   • RDocumentation
   • Stack Exchange/Overflow

Summary

• There are 5 atomic data types in R. Logical, integer, numeric, complex, character.
• Vectors and matrices are used to store data of the same type.
• Lists and data frames are used to store data of different types.
• R can handle infinity (Inf), missing values (NA), and division by 0 (NaN)
• Different data types inside of a vector or matrix will be coerced to the most flexible.

Up Next

• Exercise 2
• Break