murmurhash VS psycopg2cffi

If you have very large dicts, you might find this hash table I wrote for spaCy helpful: https://github.com/explosion/preshed . You need to key the data with 64-bit keys. We use this wrapper around murmurhash for it: https://github.com/explosion/murmurhash

There's no docs so obviously this might not be for you. But the software does work, and is efficient. It's been executed many many millions of times now.

Some notes: We don't need Sha256 and not evey base64; nothing will happen if keys will not distribute very equally. we could take MMH3; googling "python murmurhash" gives 2 interesting results; and since both use the same cpp code, let's take the one with most stars Other options would be to simply do (% NUM_SHARDS) or even shift right (however must have shards count == power of 2).

The only compatibility issue I've run into is database drivers.

For PostgreSQL, psycopg2 is not supported. psycopg2cffi is largely unmaintained, and the 2.9.0 version in PyPI lacks some newer features of psycopg2: the `psycopg2.sql` module and empty result sets raise a RuntimeError in Python 3.7+. The latest commit in on Github does have these changes [1]. Psycopg 3 [2] and pg8000 [3] (as user tlocke mentioned elsewhere) are viable alternates provided you aren't stuck with older versions of PostgreSQL. I'm going to continue to use psycopg2cffi until I can upgrade an old PostgreSQL 9.4 database.

For Microsoft SQL Server, pymssql does not support PyPy [4]. It's under new maintainership so it might gain support in the future. pypyodbc hasn't had any activity since 2022, and no new PyPI release since 2021 [5]. The datatypes returned can differ between libodbc1 versions. On Ubuntu 18.04 in particular: empty string columns are returned as a single space, integer columns are returned as a Decimal. Also, if you encounter a mysterious HY010 error ("Function sequence error"), you may need to upgrade libodbc1 to v2.3.7+ from v2.3.4 using the Microsoft repos.

[1]: https://github.com/chtd/psycopg2cffi

From time to time, I use pgcopy coupled with psycopg2cffi to feed large volumes of data processed by custom parsers written in Python for several formats. The whole process is 4-5x faster with PyPy.