cnamedtuple 0.1.1

collections.namedtuple implemented in c.

An implementation of namedtuple written in c for warp speed.

Tested against Python3.4.2 and Python3.5 0a0.

Warp Speed

Time to go fast.

Setup

Use qualified imports so that we can tell which namedtuple we are using.

>>> from namedtuple import namedtuple as cnamedtuple
>>> from collections import namedtuple as stdnamedtuple
>>> from pickle import dumps, loads

Standard Library

Running: "stdnamedtuple('std_nt_type', 'a b c d e f')"
1000 loops, best of 3: 769 usec per loop

Running: "stdnamedtuple('std_nt_type', ('a', 'b', 'c', 'd', 'e', 'f'))"
1000 loops, best of 3: 771 usec per loop

Running: "std_nt_type(1, 2, 3, 4, 5, 6)"
1000000 loops, best of 3: 1.04 usec per loop

Running: "std_nt_inst.c"
10000000 loops, best of 3: 0.12 usec per loop

Running: "std_nt_inst[3]"
10000000 loops, best of 3: 0.0532 usec per loop

Running: "std_nt_inst._replace(a=2, b=3, c=4, d=5, e=6, f=7)"
100000 loops, best of 3: 3.19 usec per loop

Running: "std_nt_inst._asdict()"
10000 loops, best of 3: 26.2 usec per loop

Running: "repr(std_nt_inst)"
1000000 loops, best of 3: 1.59 usec per loop

Running: "dumps(std_nt_inst)"
100000 loops, best of 3: 7.74 usec per loop

Running: "loads(std_nt_inst_dumped)"
100000 loops, best of 3: 4.46 usec per loop

This Implementation

Running: "cnamedtuple('c_nt_type', 'a b c d e f')"
100000 loops, best of 3: 18.8 usec per loop

Running: "cnamedtuple('c_nt_type', ('a', 'b', 'c', 'd', 'e', 'f'))"
100000 loops, best of 3: 18.5 usec per loop

Running: "c_nt_type(1, 2, 3, 4, 5, 6)"
1000000 loops, best of 3: 0.709 usec per loop

Running: "c_nt_inst.c"
10000000 loops, best of 3: 0.0657 usec per loop

Running: "c_nt_inst[3]"
10000000 loops, best of 3: 0.0523 usec per loop

Running: "c_nt_inst._replace(a=2, b=3, c=4, d=5, e=6, f=7)"
1000000 loops, best of 3: 1.85 usec per loop

Running: "c_nt_inst._asdict()"
10000 loops, best of 3: 25.9 usec per loop

Running: "repr(c_nt_inst)"
1000000 loops, best of 3: 1.32 usec per loop

Running: "dumps(c_nt_inst)"
100000 loops, best of 3: 5.72 usec per loop

Running: "loads(c_nt_inst_dumped)"
100000 loops, best of 3: 3.82 usec per loop

Contributing

The project is hosted on github.

Before submitting a patch, please make sure your Python code is PEP8 compliant and your c code is PEP7 compliant.

Contact

Please file all bug reports on github.

For questions or comments, feel free to email me at [e-mail]