clip Method¶

The clip() method constrains values to a specified range.

Mental Model

clip(lower, upper) is a guardrail: any value below lower is raised to lower, any value above upper is capped at upper, and everything in between is untouched. It is the vectorized equivalent of max(lower, min(x, upper)) applied to every element at once.

Basic Usage¶

Clipping with Lower and Upper Bounds¶

```python import pandas as pd

s = pd.Series([1, 5, 10, 15, 20, 25])

Clip values to range [5, 20]¶

clipped = s.clip(lower=5, upper=20) print(clipped) ```

0 5 1 5 2 10 3 15 4 20 5 20 dtype: int64

Only Lower or Upper Bound¶

python s.clip(lower=10) # Values below 10 become 10 s.clip(upper=15) # Values above 15 become 15

DataFrame Clipping¶

Clip All Columns¶

```python df = pd.DataFrame({ 'A': [1, 5, 10, 15, 20], 'B': [2, 8, 14, 22, 30] })

clipped = df.clip(lower=5, upper=20) print(clipped) ```

Per-Column Clipping¶

```python df = pd.DataFrame({'A': [1, 50, 100], 'B': [10, 500, 1000]})

Different bounds for each column¶

clipped = df.clip(lower=[0, 100], upper=[80, 800], axis=1) ```

Practical Examples¶

1. Outlier Capping¶

python def cap_outliers(df, columns): """Cap outliers at 1st and 99th percentiles.""" df_capped = df.copy() for col in columns: lower = df[col].quantile(0.01) upper = df[col].quantile(0.99) df_capped[col] = df[col].clip(lower=lower, upper=upper) return df_capped

2. Normalizing to Range¶

```python

Clip values to valid range before processing¶

scores = pd.Series([95, 105, 88, -5, 72]) valid_scores = scores.clip(lower=0, upper=100) ```

3. Financial Position Limits¶

```python positions = pd.DataFrame({ 'ticker': ['AAPL', 'GOOGL', 'MSFT'], 'weight': [0.45, 0.35, 0.20] })

Enforce max 40% position limit¶

positions['weight_capped'] = positions['weight'].clip(upper=0.40) ```

4. Signal Processing¶

```python

Clip noisy sensor readings¶

readings = pd.Series([98, 102, 150, 95, 99, -10, 101]) clean_readings = readings.clip(lower=90, upper=110) ```

Key Parameters¶

Parameter	Description	Default
`lower`	Minimum allowed value	None
`upper`	Maximum allowed value	None
`axis`	Align bounds along axis (0 or 1)	None
`inplace`	Modify in place	False

Notes¶

At least one of lower or upper must be specified
NaN values are not clipped (they remain NaN)
Works with both Series and DataFrame
Original data is not modified unless inplace=True

Exercises¶

Exercise 1. Create a Series with values ranging from -10 to 10. Use .clip(lower=0, upper=5) to constrain all values to the range [0, 5]. Verify that no values fall outside this range.

Solution to Exercise 1

Clip a Series to a specific range.

import pandas as pd

s = pd.Series(range(-10, 11))
clipped = s.clip(lower=0, upper=5)
print(clipped)
assert clipped.min() >= 0
assert clipped.max() <= 5
print("All values within [0, 5].")

Exercise 2. Create a DataFrame with two numeric columns. Use .clip(lower=0) to replace all negative values with 0 (no upper bound). Count how many values were clipped by comparing with the original.

Solution to Exercise 2

Clip negative values to zero and count changes.

import pandas as pd
import numpy as np

df = pd.DataFrame({
    'A': [1, -3, 5, -7, 2],
    'B': [-1, 4, -6, 8, -2]
})
clipped = df.clip(lower=0)
n_clipped = (df < 0).sum().sum()
print(clipped)
print(f"Values clipped: {n_clipped}")

Exercise 3. Given a DataFrame of daily stock returns (some extreme outliers), use .clip() to cap returns at +/- 5%. Print the number of values that were clipped on each side (above 5% and below -5%).

Solution to Exercise 3

Cap extreme stock returns at +/- 5%.

import pandas as pd
import numpy as np

np.random.seed(42)
returns = pd.DataFrame({
    'AAPL': np.random.randn(100) * 0.03,
    'MSFT': np.random.randn(100) * 0.03
})
above = (returns > 0.05).sum().sum()
below = (returns < -0.05).sum().sum()
clipped = returns.clip(lower=-0.05, upper=0.05)
print(f"Clipped above 5%: {above}")
print(f"Clipped below -5%: {below}")
print(clipped.describe())