drop Method¶

The drop() method removes specified rows or columns from a DataFrame.

Drop Columns¶

Remove columns by name.

1. Single Column¶

```python import pandas as pd

df = pd.DataFrame({ 'A': [1, 2, 3], 'B': [4, 5, 6], 'C': [7, 8, 9] })

result = df.drop('B', axis=1) print(result) ```

A C 0 1 7 1 2 8 2 3 9

2. Multiple Columns¶

python result = df.drop(['B', 'C'], axis=1)

3. columns Parameter¶

```python result = df.drop(columns=['B', 'C'])

More explicit than axis=1¶

```

Drop Rows¶

Remove rows by index.

1. Single Row¶

python result = df.drop(0) # axis=0 is default print(result)

A B C 1 2 5 8 2 3 6 9

2. Multiple Rows¶

python result = df.drop([0, 2])

3. index Parameter¶

python result = df.drop(index=[0, 2])

inplace Parameter¶

Modify DataFrame directly.

1. Without inplace¶

```python result = df.drop(columns=['B'])

df unchanged, result has change¶

```

2. With inplace¶

```python df.drop(columns=['B'], inplace=True)

df modified directly, returns None¶

```

3. Best Practice¶

```python

Prefer reassignment over inplace¶

df = df.drop(columns=['B']) ```

LeetCode Example: Tree Node¶

Drop unnecessary columns.

1. Sample Data¶

python tree = pd.DataFrame({ 'id': [1, 2, 3, 4, 5], 'p_id': [None, 1, 1, 2, 2], 'type': ['Root', 'Inner', 'Inner', 'Leaf', 'Leaf'] })

2. Drop Column¶

python result = tree.drop(columns='p_id') print(result)

id type 0 1 Root 1 2 Inner 2 3 Inner 3 4 Leaf 4 5 Leaf

3. axis=1 Syntax¶

python result = tree.drop('p_id', axis=1)

LeetCode Example: Top Travellers¶

Drop ID column from rides.

1. Sample Data¶

python rides = pd.DataFrame({ 'id': [1, 2, 3, 4], 'user_id': [1, 1, 2, 3], 'distance': [10, 15, 20, 25] })

2. Drop ID¶

python result = rides.drop('id', axis=1)

3. Result¶

python print(result)

user_id distance 0 1 10 1 1 15 2 2 20 3 3 25

Drop by Condition¶

Remove rows based on conditions.

1. Drop NaN Rows¶

```python

Use dropna instead¶

df = df.dropna() ```

2. Drop Specific Values¶

```python

Filter instead of drop¶

df = df[df['status'] != 'Deleted'] ```

3. Get Index Then Drop¶

```python

Find rows to drop¶

to_drop = df[df['value'] < 0].index df = df.drop(to_drop) ```

errors Parameter¶

Handle missing labels.

1. Default (raise)¶

```python

Raises KeyError if column doesn't exist¶

df.drop(columns=['NonExistent']) # Error! ```

2. Ignore Errors¶

```python df.drop(columns=['NonExistent'], errors='ignore')

No error, returns df unchanged¶

```

3. Safe Drop¶

```python

Drop if exists¶

columns_to_drop = ['B', 'NonExistent'] df.drop(columns=columns_to_drop, errors='ignore') ```

Drop Duplicates Context¶

Use drop_duplicates for duplicate removal.

1. drop vs drop_duplicates¶

```python

drop: remove by label¶

df.drop(index=[0, 1])

drop_duplicates: remove duplicate rows¶

df.drop_duplicates() ```

2. Different Purposes¶

```python

drop: known indices/columns¶

drop_duplicates: based on values¶

```

3. See drop_duplicates¶

Refer to drop_duplicates.md for duplicate removal.

Method Chaining¶

drop in pipelines.

1. Chain Operations¶

python result = ( df .drop(columns=['temp_col']) .drop(index=[0]) .reset_index(drop=True) )

2. With Other Methods¶

python result = ( df .assign(calculated=df['a'] + df['b']) .drop(columns=['a', 'b']) .rename(columns={'calculated': 'sum'}) )

3. Clean Pipeline¶

python result = ( raw_df .dropna() .drop(columns=['unnecessary_col']) .reset_index(drop=True) )

Exercises¶

Exercise 1. Create a DataFrame with 5 columns. Drop two columns using drop(columns=[...]). Verify the resulting DataFrame has 3 columns.

import pandas as pd

df = pd.DataFrame({
    'A': [1], 'B': [2], 'C': [3], 'D': [4], 'E': [5]
})
result = df.drop(columns=['B', 'D'])
print(result.columns.tolist())
assert len(result.columns) == 3

Exercise 2. Create a DataFrame with rows indexed [0, 1, 2, 3, 4]. Drop rows at indices 1 and 3 using drop([1, 3]). Verify the remaining indices are [0, 2, 4].

import pandas as pd

df = pd.DataFrame({'val': [10, 20, 30, 40, 50]})
result = df.drop([1, 3])
print(result)
assert result.index.tolist() == [0, 2, 4]

Exercise 3. Create a DataFrame and attempt to drop a non-existent column. Use the errors='ignore' parameter to suppress the KeyError. Then drop the column without errors='ignore' to see the error.

import pandas as pd

df = pd.DataFrame({'A': [1], 'B': [2]})
result = df.drop(columns=['C'], errors='ignore')
print(result)  # No error, df unchanged

try:
    df.drop(columns=['C'])
except KeyError as e:
    print(f"KeyError: {e}")