> [!NOTE] Index Only Scan > <table> > <tr> > <td width="25%"><img src="assets/ex_indexonlyscan.png"></td> > <td>An advanced optimization that retrieves all required columns directly from the index. By avoiding the 'Heap' (table files) entirely, it significantly reduces I/O. This requires the index to contain all requested columns and the Visibility Map to confirm the pages are 'all-visible'.</td> > </tr> > </table> > > ```sql > -- Reading a visible range entirely from the primary-key index > SET enable_seqscan = off; > > EXPLAIN (ANALYZE, COSTS, BUFFERS, VERBOSE) > SELECT id FROM animals WHERE id <= 500000; > ``` > >  > > ```text > Index Only Scan using animals_pkey on public.animals (cost=0.43..13838.40 rows=486055 width=4) (actual time=0.090..34.963 rows=500000.00 loops=1) > Output: id > Index Cond: (animals.id <= 500000) > Heap Fetches: 0 > Index Searches: 1 > Buffers: shared read=1370 > I/O Timings: shared read=8.128 > Query Identifier: 3962809790011388872 > Planning: > Buffers: shared hit=74 read=4 > I/O Timings: shared read=0.064 > Planning Time: 0.221 ms > Execution Time: 47.541 ms > ``` > >  > > <table> > <tr> > <td rowspan="2" width="25%"><img src="assets/ex_index_only_scan.svg"></td> > <td><b>Performance</b></td><td>Extremely efficient as it avoids random heap I/O.</td> > </tr> > <tr><td><b>Cost</b></td><td><code>index cost + cpu_index_tuple_cost * number of index entries</code></td></tr> > </table>