mapply

0.2.0 verified Mon Apr 27 auth: no python

Sensible multi-core apply function for Pandas. Currently version 0.2.0, supporting Pandas v3 (last version supporting Pandas v2 is 0.1.31). Active development with frequent releases.

pip install mapply

Common errors

error AttributeError: 'DataFrame' object has no attribute 'mapply' ↓

cause mapply not imported or imported after DataFrame already created.

fix

Add 'import mapply' at the top of your script, before any pandas DataFrame creation.

error ModuleNotFoundError: No module named 'mapply' ↓

cause mapply is not installed.

fix

Run: pip install mapply

error ValueError: n_workers must be >= 1 or -1 for all CPUs ↓

cause Passing invalid value to mapply.init(n_workers=...).

fix

Use n_workers=-1 for all CPUs, or an integer >= 1.

Warnings

breaking mapply 0.2.0 drops support for Pandas v2. Pandas v3+ is required. Last compatible version is 0.1.31. ↓

fix Pin mapply<0.2.0 if using Pandas v2, or upgrade to Pandas v3.

deprecated mapply 0.1.24+ declares pandas as a dependency. Older versions lack this, causing resolvers to pair with incompatible pandas versions. ↓

fix Always pin mapply>=0.1.24.

gotcha mapply must be imported before any pandas operations. Importing after creating DataFrames will not apply the monkey-patch. ↓

fix Always import mapply at the top of your script, before any pandas calls.

gotcha By default mapply uses multiprocessing (Pool). For multithreading, set POOL_CLASS to ThreadPool via mapply.init(pool_class='ThreadPool'). ↓

fix Call mapply.init(n_workers=..., pool_class='ThreadPool') for I/O-bound tasks.

Imports

mapply
wrong
```
from mapply import mapply
```
correct
```
import mapply
```
mapply is a top-level module that monkey-patches pandas; you import it, not a specific class.

Quickstart

Initialize mapply with all CPUs, then use .mapply() like .apply() but parallel.

import pandas as pd
import mapply

mapply.init(n_workers=-1)  # use all CPUs

df = pd.DataFrame({'a': [1, 2, 3], 'b': [4, 5, 6]})
result = df.mapply(lambda row: row['a'] + row['b'], axis=1)
print(result)