Forward fill and backward fill in Pandas - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

What is the output of forward fill on this DataFrame?

Given the DataFrame below, what is the result after applying df.ffill()?

Pandas

import pandas as pd
import numpy as np
df = pd.DataFrame({'A': [1, np.nan, np.nan, 4, np.nan], 'B': [np.nan, 2, np.nan, np.nan, 5]})
result = df.ffill()

A {'A': [1.0, 1.0, 1.0, 4.0, 4.0], 'B': [nan, 2.0, 2.0, 2.0, 5.0]}

B {'A': [1.0, nan, nan, 4.0, nan], 'B': [nan, 2.0, nan, nan, 5.0]}

C {'A': [1.0, 1.0, 1.0, 4.0, nan], 'B': [nan, 2.0, 2.0, 2.0, 5.0]}

D {'A': [1.0, 1.0, 1.0, 4.0, 4.0], 'B': [nan, 2.0, nan, nan, 5.0]}

💡 Hint

Forward fill replaces NaN with the last known non-NaN value going down the column.

Forward fill fills missing values by propagating the last valid observation forward. So NaNs after a number get replaced by that number until a new number appears.

❓ Predict Output

intermediate

What is the output of backward fill on this DataFrame?

Given the DataFrame below, what is the result after applying df.bfill()?

Pandas

import pandas as pd
import numpy as np
df = pd.DataFrame({'X': [np.nan, 3, np.nan, 7, np.nan], 'Y': [1, np.nan, np.nan, 4, 5]})
result = df.bfill()

A {'X': [3.0, 3.0, 7.0, 7.0, 7.0], 'Y': [1.0, 4.0, 4.0, 4.0, 5.0]}

B {'X': [3.0, 3.0, 7.0, 7.0, nan], 'Y': [1.0, 4.0, 4.0, 4.0, 5.0]}

C {'X': [nan, 3.0, 7.0, 7.0, nan], 'Y': [1.0, nan, nan, 4.0, 5.0]}

D {'X': [nan, 3.0, nan, 7.0, nan], 'Y': [1.0, nan, nan, 4.0, 5.0]}

💡 Hint

Backward fill replaces NaN with the next known non-NaN value going down the column.

Backward fill fills missing values by propagating the next valid observation backward. So NaNs before a number get replaced by that number until a new number appears.

❓ data_output

advanced

How many missing values remain after forward fill with limit?

Given the DataFrame below, after applying df.ffill(limit=1), how many NaN values remain?

Pandas

import pandas as pd
import numpy as np
df = pd.DataFrame({'A': [np.nan, np.nan, 2, np.nan, np.nan, 5]})
result = df.ffill(limit=1)

A 0

B 2

C 1

D 3

💡 Hint

Limit controls how many consecutive NaNs get filled forward.

With limit=1, only one NaN after a valid value is filled. The first two NaNs remain because no previous value exists. The NaN after 2 is filled once, but the next NaN remains.

❓ visualization

advanced

Which plot correctly shows the effect of forward fill on missing data?

You have a time series with missing values. Which plot below correctly shows the original data and the data after forward fill?

Pandas

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

dates = pd.date_range('2024-01-01', periods=6)
values = [1, np.nan, np.nan, 4, np.nan, 6]
df = pd.DataFrame({'Value': values}, index=dates)
df_ffill = df.ffill()

plt.figure(figsize=(8,4))
plt.plot(df.index, df['Value'], 'o-', label='Original')
plt.plot(df_ffill.index, df_ffill['Value'], 's--', label='Forward Fill')
plt.legend()
plt.title('Forward Fill Effect')
plt.xlabel('Date')
plt.ylabel('Value')
plt.grid(True)
plt.tight_layout()
plt.show()

A Plot with original line having gaps at NaNs and forward fill line connecting those gaps with previous values

B Plot with original line continuous and forward fill line with gaps at NaNs

C Plot with both lines continuous and identical

D Plot with original line flat and forward fill line increasing linearly

💡 Hint

Forward fill fills gaps by extending last known value, so forward fill line has no gaps.

Original data has gaps where NaNs are. Forward fill replaces NaNs with last known value, so line is continuous without gaps at those points.

🧠 Conceptual

expert

What happens if you apply forward fill then backward fill on a DataFrame with missing values?

Consider a DataFrame with missing values scattered. If you apply df.ffill() followed by df.bfill(), what is the resulting effect on missing data?

A All missing values are filled except those at the start and end of the DataFrame

B Only missing values between two valid values are filled; missing values at edges remain

C All missing values are filled because forward fill fills down and backward fill fills up

D No missing values are filled because the two methods cancel each other out

💡 Hint

Forward fill fills down, backward fill fills up, so combined they fill all gaps.

Forward fill fills missing values by propagating last valid value downward. Backward fill fills missing values by propagating next valid value upward. Together, they fill all missing values except if entire column is NaN.