simon987/pg_hamming
1
0
Fork 0
mirror of https://github.com/simon987/pg_hamming.git synced 2025-04-19 18:26:43 +00:00
Code Issues Projects Releases Wiki Activity

fix + optimization

Browse Source
This commit is contained in:
simon 2019-07-20 12:51:08 -04:00
parent 1dc4edad8a
commit 3cf7882cb6
1 changed files with 30 additions and 28 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

52
hamming.c
View File

@ -3,18 +3,13 @@
PG_MODULE_MAGIC; PG_MODULE_MAGIC;
int const HASH_SIZE = 12;
int const HASH_BITS = HASH_SIZE * HASH_SIZE;
int const BYTEA_LEN = HASH_BITS / 8;
PG_FUNCTION_INFO_V1(hash_is_within_distance); PG_FUNCTION_INFO_V1(hash_is_within_distance);
/** /**
* Check if the hamming distance of the two raw byte arrays * Check if the hamming distance of the two raw byte arrays
* is within the specified distance * is within the specified distance
* *
* It is assumed that: the two arrays are exactly * It is assumed that: the two arrays are exactly 18 bytes long
* BYTEA_LEN bytes long
* *
* Import with * Import with
CREATE OR REPLACE FUNCTION hash_is_within_distance(bytea, bytea, integer) RETURNS boolean CREATE OR REPLACE FUNCTION hash_is_within_distance(bytea, bytea, integer) RETURNS boolean
@ -25,24 +20,29 @@ PG_FUNCTION_INFO_V1(hash_is_within_distance);
*/ */
Datum hash_is_within_distance(PG_FUNCTION_ARGS) { Datum hash_is_within_distance(PG_FUNCTION_ARGS) {
bytea *hash1 = PG_GETARG_BYTEA_P(0); char *h1 = VARDATA(PG_GETARG_BYTEA_P(0));
bytea *hash2 = PG_GETARG_BYTEA_P(1); char *h2 = VARDATA(PG_GETARG_BYTEA_P(1));
int32 max_distance = PG_GETARG_INT32(2); int32 max_distance = PG_GETARG_INT32(2);
int distance = 0; int distance = 0;
char *h1 = hash1->vl_len_; distance += __builtin_popcountll(
char *h2 = hash2->vl_len_; *((uint64 *) h1) ^ *((uint64 *) h2)
);
for (int i = BYTEA_LEN; i >= 0; i--) {
distance += __builtin_popcount(h1[i] ^ h2[i]);
if (distance > max_distance) { if (distance > max_distance) {
PG_RETURN_BOOL(false); PG_RETURN_BOOL(false);
} }
distance += __builtin_popcountll(
*((uint64 *) h1 + 1) ^ *((uint64 *) h2 + 1)
);
if (distance > max_distance) {
PG_RETURN_BOOL(false);
} }
distance += __builtin_popcount(
*((uint16 *) h1 + 8) ^ *((uint16 *) h2 + 8)
);
PG_RETURN_BOOL(true); PG_RETURN_BOOL(distance <= max_distance);
} }
PG_FUNCTION_INFO_V1(hash_distance); PG_FUNCTION_INFO_V1(hash_distance);
@ -50,8 +50,7 @@ PG_FUNCTION_INFO_V1(hash_distance);
/** /**
* Hamming distance of two raw byte arrays * Hamming distance of two raw byte arrays
* *
* It is assumed that: the two arrays are exactly * It is assumed that: the two arrays are exactly 18 bytes long
* BYTEA_LEN bytes long
* *
* Import with * Import with
CREATE OR REPLACE FUNCTION hash_distance(bytea, bytea) RETURNS integer CREATE OR REPLACE FUNCTION hash_distance(bytea, bytea) RETURNS integer
@ -62,17 +61,20 @@ PG_FUNCTION_INFO_V1(hash_distance);
*/ */
Datum hash_distance(PG_FUNCTION_ARGS) { Datum hash_distance(PG_FUNCTION_ARGS) {
bytea *hash1 = PG_GETARG_BYTEA_P(0); char *h1 = VARDATA(PG_GETARG_BYTEA_P(0));
bytea *hash2 = PG_GETARG_BYTEA_P(1); char *h2 = VARDATA(PG_GETARG_BYTEA_P(1));
int distance = 0; int distance = 0;
char *h1 = hash1->vl_len_; distance += __builtin_popcountll(
char *h2 = hash2->vl_len_; *((uint64 *) h1) ^ *((uint64 *) h2)
);
for (int i = BYTEA_LEN; i >= 0; i--) { distance += __builtin_popcountll(
distance += __builtin_popcount(h1[i] ^ h2[i]); *((uint64 *) h1 + 1) ^ *((uint64 *) h2 + 1)
} );
distance += __builtin_popcount(
*((uint16 *) h1 + 8) ^ *((uint16 *) h2 + 8)
);
PG_RETURN_INT32(distance); PG_RETURN_INT32(distance);
} }