INNER JOIN performance depends on indexes on join columns (foreign keys and primary keys), table sizes (smaller tables join faster), number of matching rows (fewer matches = faster), and WHERE clause filtering (filter early when possible).

Index ALL join columns - most critical optimization (100x speedup possible)

Create index on foreign keys: CREATE INDEX idx_orders_customer ON orders(customer_id)

Primary keys usually auto-indexed, but verify with SHOW INDEXES

Use covering indexes that include all SELECT columns for index-only scans

Filter with WHERE to reduce rows before or after join

Avoid functions on join columns: LOWER(c.id) prevents index usage

Select specific columns, not SELECT * to reduce data transfer

Keep table statistics updated: ANALYZE TABLE for optimal query plans

Smaller table × larger table usually faster (optimizer handles this)

Use EXPLAIN ANALYZE to identify missing indexes and slow operations

Forgetting ON clause creates cartesian product (every row × every row)

Not indexing foreign key columns causes full table scans

Using old comma syntax (FROM a, b WHERE) instead of explicit INNER JOIN

Confusing INNER JOIN (excludes non-matches) with LEFT JOIN (preserves left table)

Putting filters in ON clause instead of WHERE (works but less clear)

Using functions on join columns: ON YEAR(o.date) = c.year prevents indexes

Expecting NULL foreign keys to match (they never do)

Not using table aliases makes queries verbose and hard to read

Selecting SELECT * when only few columns needed wastes resources

Assuming join order matters (optimizer chooses, but indexes matter more)