apply, map and vectorizing in Python

apply is a loop wearing a nice hat: reach for a vectorized expression first.

import pandas as pd

frame = pd.DataFrame({"chars": [420, 380], "errors": [3, 0]})

print(frame["chars"].map(lambda n: n // 5))
print(frame.apply(lambda col: col.max() - col.min()))
print(frame.apply(lambda row: row["chars"] - row["errors"], axis=1))

# the same thing, vectorized
print((frame["chars"] - frame["errors"]).tolist())

How it works

1. Series.map applies elementwise, DataFrame.apply per column or row.
2. axis=1 walks rows, which is the slow direction.
3. Arithmetic on whole columns is what pandas is fast at.

Keywords and builtins used here

• as
• lambda
• print

The run, in numbers

Lines

10

Characters to type

331

Tokens

125

Three-star pace

105 tpm

At the three-star pace of 105 tokens a minute, this run takes about 71 seconds.

Type this snippet

Step 2 of 5 in Deriving columns, step 10 of 26 in pandas.

← Previous Next →