Which PySpark method returns a DataFrame containing the first N rows?

Which PySpark method returns a DataFrame containing the first N rows?

• A. limit(N)
• B. take(N)
• C. head(N)
• D. collect()[0:N]

Answer: (A)

Limiting the DataFrame to the first N rows is done with a transformation that returns another DataFrame containing those rows. Use limit(N) to get a DataFrame with at most N rows. This is the best choice because it preserves the DataFrame type and avoids pulling data to the driver. The other options return Python collections rather than a DataFrame: take(N) and head(N) yield a list of Row objects; collect() gathers all rows to the driver, and slicing after collect (collect()[0:N]) also gives a Python list instead of a DataFrame. If you need a deterministic order before limiting, apply an orderBy first, since without explicit ordering the notion of “first” isn’t guaranteed.

Limiting the DataFrame to the first N rows is done with a transformation that returns another DataFrame containing those rows. Use limit(N) to get a DataFrame with at most N rows. This is the best choice because it preserves the DataFrame type and avoids pulling data to the driver.

The other options return Python collections rather than a DataFrame: take(N) and head(N) yield a list of Row objects; collect() gathers all rows to the driver, and slicing after collect (collect()[0:N]) also gives a Python list instead of a DataFrame. If you need a deterministic order before limiting, apply an orderBy first, since without explicit ordering the notion of “first” isn’t guaranteed.