Perl 🐪 map versus Python 🐍 dict Comprehension
Perl’s map function always returns lists, even when the intent to create a
hash is abundantly clear. This leads to CPU cycles wasted on creating the hash
from the list after-the-fact.
Python’s dict comprehension directly constructs a dictionary, the Python
equivalent to a Perl hash, on the fly, without forcing an intermediate list
stage.
Results
tl;dr
Python wins by taking about half the time Perl takes.
Stats for Nerds
Perl
| Test | Time Taken (s) |
|---|---|
| Regular Hash | 0.452912092208862 |
| Keys Assignment | 0.452517986297607 |
| Numeric Regular Hash | 2.83940291404724 |
| Numeric Keys Assignment | 2.83703708648682 |
| Numeric No Hash | 0.647165060043335 |
Python
| Test | Time Taken (s) |
|---|---|
| Regular Hash | 0.23767730799954734 |
| Numeric Regular Hash | 0.8262694709992502 |
Perl with Loops (not map)
| Test | Time Taken (s) |
|---|---|
| Regular Hash | 0.336189031600952 |
| Keys Assignment | 0.339548110961914 |
| Numeric Regular Hash | 1.44092202186584 |
| Numeric Keys Assignment | 1.47567176818848 |
PyPy
| Test | Time Taken (s) |
|---|---|
| Regular Hash | 0.06730515199888032 |
| Numeric Regular Hash | 0.30881453300025896 |
Why?
This Twitter thread sparked a discussion leading to this benchmarking code being written.
How?
git clone https://gitlab.com/ankitpati/pl-map-vs-py-dcomp.git
cd pl-map-vs-py-dcomp/
./run.shFor tweaking things, consult the source!