SPARK Reference Guide
Revision Time: 4 mins

Join Reference Reference

Relational join patterns, conditions, and broadcast helpers.

join
Return: DataFrame

Joins with another DataFrame on join columns or expressions.

Syntax signature:df1.join(df2, on, how='inner')
Code snippet:
python
df1.join(df2, on='id', how='left')
Remember: Specify explicit left/right conditions with syntax: (df1.id == df2.id) to resolve duplicate column name bugs.
broadcast
Return: DataFrame

Marks a smaller DataFrame to be broadcasted to all executors.

Syntax signature:broadcast(df)
Code snippet:
python
df1.join(broadcast(df2), on='key')
Remember: Forces a Map-Side Broadcast Join which keeps the big dataset stationary, bypassing shuffles completely.