Reading and writing in one operation (high-level)

The generic swap combines x[i] and x[i] <- value in a single operation.

Usage

swap(x, value, ...)
# S3 method for class 'ff'
swap(x, value, i, add = FALSE, pack = FALSE, ...)
# S3 method for class 'ff_array'
swap(x, value, ..., bydim = NULL, drop = getOption("ffdrop"), add = FALSE, pack = FALSE)
# Default S3 method
swap(x, value, ..., add = FALSE)

Arguments

x: a ff or ram object
value: the new values to write, possibly recycled, see [.ff
i: index information, see [.ff
...: missing OR up to length(dim(x)) index expressions OR (ff only) hi objects
drop: logical scalar indicating whether array dimensions shall be dropped
bydim: how to interpret vector to array data, see [.ff
add: TRUE if the values should rather increment than overwrite at the target positions, see readwrite.ff
pack: FALSE to prevent rle-packing in hybrid index preprocessing, see as.hi

Details


  y <- swap(x, value, i, add=FALSE, ...)

  is a shorter and more efficient version of

  y <- x[i, add=FALSE, ...]
  x[i, add=FALSE, ...] <- value

  and

  y <- swap(x, value, i, add=TRUE, ...)

  is a shorter and more efficient version of

  y <- x[i, add=TRUE, ...]
  y <- y + value
  x[i, add=FALSE, ...] <- y

Value

Values at the target positions. More precisely swap(x, value, i, add=FALSE) returns the old values at the position i while swap(x, value, i, add=TRUE) returns the incremented values of x.

Author

Jens Oehlschlägel

Note

Note that swap.default changes the object in its parent frame and thus violates R's usual functional programming logic. When using add=TRUE, duplicated index positions should be avoided, because ff and ram objects behave differently:


  swap.ff(x, 1, c(3,3), add=TRUE)
  # will increment x at position 3 TWICE by 1, while
  swap.default(x, 1, c(3,3), add=TRUE)
  # will increment x at position 3 just ONCE by 1

Examples

  x <- ff("a", levels=letters, length=52)
  y <- swap(x, "b", sample(length(x), 26))
  x
#> ff (open) integer length=52 (52) levels: a b c d e f g h i j k l m n o p q r s t u v w x y z
#>  [1]  [2]  [3]  [4]  [5]  [6]  [7]  [8]      [45] [46] [47] [48] [49] [50] [51] 
#>    a    b    b    a    a    a    a    a    :    a    b    a    b    a    a    b 
#> [52] 
#>    b 
  y
#>  [1] a a a a a a a a a a a a a a a a a a a a a a a a a a
#> Levels: a b c d e f g h i j k l m n o p q r s t u v w x y z
  rm(x,y); gc()
#>           used (Mb) gc trigger  (Mb) max used  (Mb)
#> Ncells 1271406 68.0    2239654 119.7  2239654 119.7
#> Vcells 2345082 17.9    8388608  64.0  8387257  64.0