Package 'bitops'

Bitwise And, Or and Xor Operations

Description

Bitwise operations, ‘and’ (&), ‘or’ (|), and ‘Xor’ (xor).

Usage

bitAnd(a, b)
a %&% b
bitOr (a, b)
a %|% b
bitXor(a, b)
a %^% b

Arguments

a, b	numeric vectors of compatible length, each treated as 32 bit “strings”.

Details

The bitwise operations are applied to the arguments cast as 32 bit (unsigned long) integers. NA is returned wherever the magnitude of the arguments is not less than $2^31$, or, where either of the arguments is not finite.

For bitwise ‘not’ (! in R), use bitFlip().

Value

non-negative integer valued numeric vector of maximum length of a or b.

Author(s)

Steve Dutky; idea for operators: Dan L Robertson

See Also

bitFlip, bitShiftL; further, cksum.

Examples

bitAnd(15,7) == 7 ;  identical(15 %&% 7, bitAnd(15, 7))
bitOr(15,7) == 15 ;  identical(15 %|% 7, bitOr (15, 7))
bitXor(15,7) == 8 ;  identical(15 %^% 7, bitXor(15,7))
bitOr(-1,0) == 4294967295 ; identical(-1 %|% 0, bitOr(-1,0))

---

Binary Flip (Not) Operator

Description

The binary flip (‘not’, R's !) operator, bitFlip(a, w), “flips every bit” of a up to the w-th bit.

Usage

bitFlip(a, bitWidth = 32)

Arguments

a	numeric vector.
bitWidth	scalar integer between 0 and 32.

Value

(“binary”) numeric vector of the same length as a masked with (2^bitWidth)-1. NA is returned for any value of a that is not finite or whose magnitude is greater or equal to $2^{32}$.

Note

bitFlip(a, w) is an “involution”, i.e. it is its own inverse – when a is in $\{0, 1, .., 2^{32}-1\}$. Notably, negative values a are equivalent to their values in the above range, see also bitUnique() in the ‘Examples’.

Author(s)

Steve Dutky

See Also

bitShiftL, bitXor, etc.

Examples

bitFlip(0:5)
##
bitUnique <- function(x) bitFlip(bitFlip(x)) # "identity" when x in  0:(2^32-1)
bitUnique(  0:16 ) # identical (well, double precision)
bitUnique(-(1:16)) # 4294967295 ...
stopifnot(
  identical(bitUnique(-(1:16)), 2^32 -(1:16)),
  bitFlip(-1) == 0,
  bitFlip(0 ) == 2^32 - 1,
  bitFlip(0, bitWidth=8) == 255
)

---

Bitwise Shift Operator (to the Left or Right)

Description

These functions shift integers bitwise to the left or to the right, returning unsigned integers, i.e., values in ${0, 1, \ldots, 2^{32}-1}$.

Usage

bitShiftL(a, b)
a %<<% b
bitShiftR(a, b)
a %>>% b

Arguments

a	numeric vector (integer valued), to be shifted.
b	integer (valued) vector. Internally, only b %% 32 is used, e.g, b = 32 is equivalent to b = 0, i.e., no shift. This corresponds to cyclic rotation (to the left or right).

Value

non-negative integer valued numeric vector of maximum length of a or b containing the value of a shifted to the left or right by b bits. NA is returned wherever the value of a or b is not finite, or, wherever the magnitude of a is greater than or equal to $2^{32}$.

See Also

bitFlip, bitXor, etc.

Examples

bitShiftL(0:4, 1) # 0 2 4 6 8
 bitShiftL(0:3, 2) # 0 4 8 12

 stopifnot(exprs = {
     identical(bitShiftL(0:4, 1), 0:4 %<<% 1)
     identical(bitShiftR(0:3, 2), 0:3 %>>% 2)
})

 bitShiftR(0:7, 1) # 0 0  1 1  2 2  3 3 <==> N %/% 2
 bitShiftR(0:7, 2) # 0 0 0 0  1 1 1 1   <==> N %/% 4
 ## all outputs are "unsigned integer" :
 stopifnot( bitShiftL(-1, 0) == 2^32 - 1   ,
            bitShiftL(-7, 0) == 4294967289 ,
            bitShiftL(-7, 0) == bitShiftR(-7, 0))

 bitShiftR(-1,1) == 2147483647
 bitShiftL(2147483647,1) == 4294967294 # <==> * 2
 bitShiftL( -1,       1) == 4294967294

 bitShiftL(47, 32) # is 47

## 5 Christmas trees  ( bitShiftL *rotates* to the left)
t(outer(1:5, 0:40, bitShiftL))

N <- as.numeric( rpois(1000, 100) )
stopifnot(identical(bitShiftL(N,0),   N),
          identical(bitShiftL(N,1), 2*N),
          identical(bitShiftL(N,2), 4*N),
          ## right shift:
          identical(bitShiftR(N,2), N %/% 4),
          identical(bitShiftR(N,4), N %/% 16))

---

Compute Check Sum

Description

Return a cyclic redundancy checksum for each element in the argument.

Usage

cksum(a)

Arguments

a	coerced to character vector

Details

NA's appearing in the argument are returned as NA's.

The default calculation is identical to that given in pseudo-code in the ACM article (in the References).

Value

numeric vector of the same length as a.

Author(s)

Steve Dutky [email protected]

References

Fashioned from cksum(1) UNIX command line utility, i.e., man cksum.

Dilip V. Sarwate (1988) Computation of Cyclic Redundancy Checks Via Table Lookup, Communications of the ACM 31, 8, 1008–1013.

See Also

bitShiftL, bitAnd, etc.

Examples

b <- "I would rather have a bottle in front of me than frontal lobotomy\n"
 stopifnot(cksum(b) == 1342168430)
 (bv <- strsplit(b, " ")[[1]])
 cksum(bv) # now a vector of length 13

---