Analyse vector sequences

How can you easily analyse vector sequences in base R?

Run-length encoding

Flying over the GitHub repository about useful and powerful shortcuts in base R, I stumbled upon an easy way to split a vector into sequences of equal values.

The function rle() allows you to deconstruct a vector. This function shows the frequency of an instance in a certain sequence. If a value is followed by the same one, the rle() function will count the number of these successive instances.

The rle() function returns a list of the values and their respective lengths in a vector sequence. The class of this list is "rle".

x <- c(1L, 5L, 7L, 2L, 2L, 7L, 7L, 7L, 8L, 1L)
(y <- rle(x))

Run Length Encoding
  lengths: int [1:7] 1 1 1 2 3 1 1
  values : int [1:7] 1 5 7 2 7 8 1

tidyr::tibble(
  length_x = y[["lengths"]], values_x = y[["values"]]
)

# A tibble: 7 × 2
  length_x values_x
     <int>    <int>
1        1        1
2        1        5
3        1        7
4        2        2
5        3        7
6        1        8
7        1        1

typeof(y)

[1] "list"

class(y)

[1] "rle"

w <- c(2.0, 3.0, 8.0, 8.0, 5.0, 5.0, 5.0, 7.0, 9.0, 9.0)
(z <- rle(w))

Run Length Encoding
  lengths: int [1:6] 1 1 2 3 1 2
  values : num [1:6] 2 3 8 5 7 9

tidyr::tibble(
  length_w = z[["lengths"]], values_w = z[["values"]]
)

# A tibble: 6 × 2
  length_w values_w
     <int>    <dbl>
1        1        2
2        1        3
3        2        8
4        3        5
5        1        7
6        2        9

typeof(z)

[1] "list"

class(z)

[1] "rle"

m <- c(FALSE, FALSE, FALSE, TRUE, TRUE, FALSE, FALSE, TRUE)
(n <- rle(m))

Run Length Encoding
  lengths: int [1:4] 3 2 2 1
  values : logi [1:4] FALSE TRUE FALSE TRUE

tidyr::tibble(
  length_m = n[["lengths"]], values_m = n[["values"]]
)

# A tibble: 4 × 2
  length_m values_m
     <int> <lgl>   
1        3 FALSE   
2        2 TRUE    
3        2 FALSE   
4        1 TRUE    

typeof(n)

[1] "list"

class(n)

[1] "rle"

Reverse operation

You can get the vector back with inverse.rle().

(inv_y <- inverse.rle(y))

 [1] 1 5 7 2 2 7 7 7 8 1

typeof(inv_y)

[1] "integer"

class(inv_y)

[1] "integer"

(inv_z <- inverse.rle(z))

 [1] 2 3 8 8 5 5 5 7 9 9

typeof(inv_z)

[1] "double"

class(inv_z)

[1] "numeric"

(inv_n <- inverse.rle(n))

[1] FALSE FALSE FALSE  TRUE  TRUE FALSE FALSE  TRUE

typeof(inv_n)

[1] "logical"

class(inv_n)

[1] "logical"