This lesson's interactive features are locked, please to keep using them

Concatenating DataFrames

We've been using pd.merge() to combine related DataFrames by matching keys. By default, pd.concat() solves a different problem: stacking them vertically. Here, we're literally just adding more records to the DataFrame.

Say you have two months of sales data with the same columns:

sales_jan = pd.DataFrame(
    {
        "sale_id": ["K8mQ2xV7", "N4pL9sR2"],
        "month": ["Jan", "Jan"],
        "amount": [100, 150],
    }
)

sales_feb = pd.DataFrame(
    {
        "sale_id": ["B7vD3kT6", "R2cF8nJ5"],
        "month": ["Feb", "Feb"],
        "amount": [200, 125],
    }
)

Concatenate them to create one taller DataFrame:

combined = pd.concat([sales_jan, sales_feb])
print(combined)
#     sale_id  month  amount
# 0  K8mQ2xV7    Jan     100
# 1  N4pL9sR2    Jan     150
# 0  B7vD3kT6    Feb     200
# 1  R2cF8nJ5    Feb     125

Ignore the Index

By default, pd.concat() keeps the original index labels from each DataFrame. That's why the result above has two rows labeled 0 and two rows labeled 1.

That's usually not what you want! Set ignore_index=True to create a fresh sequential index:

combined = pd.concat([sales_jan, sales_feb], ignore_index=True)
print(combined)
#     sale_id  month  amount
# 0  K8mQ2xV7    Jan     100
# 1  N4pL9sR2    Jan     150
# 2  B7vD3kT6    Feb     200
# 3  R2cF8nJ5    Feb     125

Real-world Pandas code is full of df.reset_index(), ignore_index=True on pd.concat(), and similar "index hygiene." This is how analysts work around Pandas' quirky default behavior.

Assignment

Complete the combine_monthly_sales function. It accepts two DataFrames: sales_jan and sales_feb. It should:

Create a copy of each DataFrame.
Add a month column containing "Jan" to the January copy and a month column containing "Feb" to the February copy.
Concatenate the rows with ignore_index=True.
Return the combined DataFrame.