Hashing a String to a Numeric Value in PostgreSQL
I need to Convert Strings stored in my Database to a Numeric value. Result can be Integer (preferred) or Bigint. This conversion is to be done at Database side in a PL/pgSQL function.
Can someone please point me to some algorithm or any API's that can be used to achieve this?
I have been searching for this on Google for hours now, could not find anything useful so far :(
Just keep the first 32 bits or 64 bits of the MD5 hash. Of course, it voids the main property of md5 (=the probability of collision being infinitesimal) but you'll still get a wide dispersion of values which presumably is good enough for your problem.
SQL functions derived from the other answers:
For bigint:
create function h_bigint(text) returns bigint as $$
select ('x'||substr(md5($1),1,16))::bit(64)::bigint;
$$ language sql;
For int:
create function h_int(text) returns int as $$
select ('x'||substr(md5($1),1,8))::bit(32)::int;
$$ language sql;
You can create a md5 hash value without problems:
select md5('hello, world');
This returns a string with a hex number.
Unfortunately there is no built-in function to convert hex to integer but as you are doing that in PL/pgSQL anyway, this might help:
https://stackoverflow.com/a/8316731/330315
Must it be an integer? The pg_crypto module provides a number of standard hash functions (md5, sha1, etc). They all return bytea. I suppose you could throw away some bits and convert bytea to integer.
bigint is too small to store a cryptographic hash. The largest non-bytea binary type Pg supports is uuid. You could cast a digest to uuid like this:
select ('{'||encode( substring(digest('foobar','sha256') from 1 for 16), 'hex')||'}')::uuid;
uuid
--------------------------------------
c3ab8ff1-3720-e8ad-9047-dd39466b3c89