hash.q

Non-cryptographic hashes - xxHash (xxh32, xxh64, xxh3_64, xxh3_128), rapidhash and MurmurHash3 - for bucketing, keying, sampling and fingerprinting (md5 and .Q.sha1 are the digests). Every function hashes the bytes of a string, a byte vector or a symbol (its text); a char atom is one byte. A symbol vector, or a list of those subjects, answers one hash per element in a single call (a list of one-char strings is spelt (enlist "a";enlist "b") - ("a";"b") is the string "ab"); anything else is 'type. A 32-bit hash rides a long; a 64-bit hash is the long with the same bit pattern (negative when the top bit is set); a 128-bit hash is 16 bytes, a list of them over a collection. The plain forms use seed 0; .hash.seeded takes the seed explicitly. The values are those of DuckDB's hashfuncs extension for the same bytes and seed, so a key bucketed in q matches one bucketed in SQL.

Examples

.hash.xxh64 "hello world"

.hash.xxh64 `AAPL`MSFT`IBM

.hash.seeded[`murmur3_32;42;"hello world"]

Entity Summary

Entities	Short Description
.hash.murmur3_128[x]	MurmurHash3 x86_128 with seed 0, as the 16 bytes the reference writes (four little-endian words), what every MurmurHash3 library hands back as bytes; `raze string reverse` reads them as one little-endian number.
.hash.murmur3_32[x]	MurmurHash3 x86_32 with seed 0, as a long holding the unsigned 32-bit value.
.hash.murmur3_x64_128[x]	MurmurHash3 x64_128 with seed 0, as the 16 bytes the reference writes (two little-endian words), what every MurmurHash3 library hands back as bytes; `raze string reverse` reads them as one little-endian number.
.hash.rapidhash[x]	rapidhash v3 with seed 0, as the long with the same 64 bits.
.hash.seeded[algo;seed;data]	Any of the eight hashes with an explicit seed; seed 0 is the plain form.
.hash.xxh32[x]	xxHash XXH32 with seed 0, as a long holding the unsigned 32-bit value.
.hash.xxh3_128[x]	xxHash XXH3 128-bit with seed 0, as 16 bytes in xxHash's canonical order (high 64 bits first), so `raze string` of the result is the hex other xxHash implementations print.
.hash.xxh3_64[x]	xxHash XXH3 64-bit with seed 0, as the long with the same 64 bits.
.hash.xxh64[x]	xxHash XXH64 with seed 0, as the long with the same 64 bits.

Entity Details

.hash.murmur3_128[x]

MurmurHash3 x86_128 with seed 0, as the 16 bytes the reference writes (four little-endian words), what every MurmurHash3 library hands back as bytes; `raze string reverse` reads them as one little-endian number.

Examples

.hash.murmur3_128 "abc"

.hash.murmur3_32[x]

MurmurHash3 x86_32 with seed 0, as a long holding the unsigned 32-bit value.

Examples

.hash.murmur3_32 "abc"

.hash.murmur3_x64_128[x]

MurmurHash3 x64_128 with seed 0, as the 16 bytes the reference writes (two little-endian words), what every MurmurHash3 library hands back as bytes; `raze string reverse` reads them as one little-endian number.

Examples

.hash.murmur3_x64_128 "abc"

.hash.rapidhash[x]

rapidhash v3 with seed 0, as the long with the same 64 bits.

Examples

.hash.rapidhash "abc"

.hash.seeded[algo;seed;data]

Any of the eight hashes with an explicit seed; seed 0 is the plain form.

Return:	as the algorithm's plain form; a vector (or list of 16-byte vectors) over a collection
Parameters:	seed - (long) any integer atom: xxh32 and the murmur3 hashes take 0 to 4294967295, the others any 64-bit pattern (-1 is the seed 18446744073709551615) data - (any) the string, byte vector, symbol or char to hash, or a symbol vector / list of those for one hash each algo - (symbol) one of `xxh32`xxh64`xxh3_64`xxh3_128`rapidhash`murmur3_32`murmur3_128`murmur3_x64_128; the same name as a string also reads
throws:	domain - an unknown algorithm, or a 32-bit seed out of range type - data (or an element of it) is not text, bytes or a symbol, or seed is not an integer

Examples

.hash.seeded[`xxh64;42;"abc"]

.hash.seeded[`murmur3_32;42;("abc";`abc;0x616263)]

.hash.xxh32[x]

xxHash XXH32 with seed 0, as a long holding the unsigned 32-bit value.

Examples

.hash.xxh32 "abc"

.hash.xxh3_128[x]

xxHash XXH3 128-bit with seed 0, as 16 bytes in xxHash's canonical order (high 64 bits first), so `raze string` of the result is the hex other xxHash implementations print.

Examples

raze string .hash.xxh3_128 "abc"

.hash.xxh3_64[x]

xxHash XXH3 64-bit with seed 0, as the long with the same 64 bits.

Examples

.hash.xxh3_64 "abc"

.hash.xxh64[x]

xxHash XXH64 with seed 0, as the long with the same 64 bits.

Examples

.hash.xxh64 "abc"