Processing nodes

Once you have a set of rows, the next type of node you'll encounter when using a single table are ones that process that set in various ways. These nodes typically take in a row set and output a different row set, of either the same size or smaller (perhaps only a single value).

Sort

Sort nodes can appear when you insert ORDER BY statements into your queries:

EXPLAIN ANALYZE SELECT customerid FROM customers ORDER BY zip;QUERY PLAN----------Sort (cost=2104.77..2154.77 rows=20000 width=8) (actual time=162.796..199.971 rows=20000 loops=1)Sort Key: zipSort Method:  external sort  Disk: 352kB->  Seq Scan on customers  (cost=0.00..676.00 rows=20000 width=8) (actual time=0.013..46.748 rows=20000 loops=1)Total runtime: 234.527 ms

Sort operations can either execute in memory using the quicksort algorithm, if they're expected to fit, or will be swapped to disk to use what's called an external merge sort—the case in this example. The threshold at which that happens depends on the work_mem...