Indexing

Indexing with Positive Integers

• there are circumstances where we want to select only some of the elements of a vector/array/dataframe/list
• this selection is done using subscripts (also known as indices)
• subscripts have square brackets [2] while functions have round brackets (2)
• Subscripts on vectors, matrices, arrays and dataframes have one set of square brackets [6], [3,4] or [2,3,2,1]

• when a subscript appears as a blank it is understood to mean all of thus

  • [,4] means all rows in column 4 of an object
  • [2,] means all columns in row 2 of an object.
  • subscripts on lists have (usually) double square brackets [ [2] ] or [ [i,j] ]

• A vector of positive integers as index:The index vector can be of any length and the result is of the same length as the index vector. For example,

   1 > letters[1:3]
   2 [1] "a" "b" "c"
   3 > letters[c(1:3,1:3)]
   4 [1] "a" "b" "c" "a" "b" "c"

• A logical vector as index: Values corresponding to T values in the index vector are selected and those corresponding to F or NA are omitted. For example,

   1 > x<-c(1,2,3,NA)
   2 > x[!is.na(x)]
   3 [1] 1 2 3

creates a vector without missing values. Also

   1 > x[is.na(x)] <- 0
   2 > x
   3 [1] 1 2 3 0

replaces the missing value by zeros. A common operation is to select rows or columns of data frame that meet some criteria. For example, to select those rows of painters data frame with Colour >= 17:

   1 > library(MASS)
   2 > painters[painters$Colour >= 17,]
   3 Composition Drawing Colour Expression School
   4 Bassano          6       8     17          0      D
   5 Giorgione        8       9     18          4      D
   6 Pordenone        8      14     17          5      D

We may want to select on more than one criterion. We can combine logical indices by the 'and', 'or' and 'not' operators. For example,

   1 > painters[painters$Colour >= 17 &  Composition Drawing Colour
   2 Titian             12      15     18
   3 Rembrandt          15       6     17
   4 Rubens             18      13     17
   5 Van Dyck           15      10     17

List of Logical Operations

If we want to select a subgroup, for example those with schools A, B, and D. We can generate a logical vector using the <latex> \mathtt{\%in\%}</latex> operator as follows:

   1 > painters[painters$School %in% c("A","C","D"),]
   2 Da Udine           10       8     16        3      A
   3 Da Vinci           15      16      4       14      A
   4 Del Piombo          8      13     16        7      A

Sometimes we are interested in the indices of rows satisfying a certain condition. To extract these indices we use the which() command.

   1 > which(painters$School %in% c("A","C","D"))
   2 [1]  1  2  3  4  5  6  7  8  9 10 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
   3 [26] 32

Indexing with Character Vectors

A vector character strings with variable names can be used to extract those variables relevant for analysis. This is very useful when we have a large number of variables and we need to work with a few ones. For example,

   1 > names(painters)
   2 [1] "Composition" "Drawing" "Colour" "Expression" "School"
   3 > painters[1:3,c("Drawing","Expression")]
   4 Drawing Expression
   5 Da Udine         8          3
   6 Da Vinci        16         14
   7 Del Piombo      13          7

• a vector of character strings could a index on a vector when the vector has names:

   1 > x <- c(1:3,NA)
   2 > names(x)<-letters[1:4]
   3 > x
   4 a  b  c  d
   5 1  2  3 NA
   6 > x[c("a","c")]
   7 a c
   8 1 3

Trimming Vectors Using Negative Indices

• an extremely useful facility is to use negative indices to drop terms from a vector
• suppose we wanted a new vector, z, to contain everything but the first element of x

   1 > x<- c(5,8,6,7,1,5,3)
   2 > (z <- x[-1])
   3 [1] 8 6 7 1 5 3