1---
2title: Batch INSERT Statements for Bulk Data
3impact: MEDIUM
4impactDescription: 10-50x faster bulk inserts
5tags: batch, insert, bulk, performance, copy
6---
7 
8## Batch INSERT Statements for Bulk Data
9 
10Individual INSERT statements have high overhead. Batch multiple rows in single statements or use COPY.
11 
12**Incorrect (individual inserts):**
13 
14```sql
15-- Each insert is a separate transaction and round trip
16insert into events (user_id, action) values (1, 'click');
17insert into events (user_id, action) values (1, 'view');
18insert into events (user_id, action) values (2, 'click');
19-- ... 1000 more individual inserts
20 
21-- 1000 inserts = 1000 round trips = slow
22```
23 
24**Correct (batch insert):**
25 
26```sql
27-- Multiple rows in single statement
28insert into events (user_id, action) values
29  (1, 'click'),
30  (1, 'view'),
31  (2, 'click'),
32  -- ... up to ~1000 rows per batch
33  (999, 'view');
34 
35-- One round trip for 1000 rows
36```
37 
38For large imports, use COPY:
39 
40```sql
41-- COPY is fastest for bulk loading
42copy events (user_id, action, created_at)
43from '/path/to/data.csv'
44with (format csv, header true);
45 
46-- Or from stdin in application
47copy events (user_id, action) from stdin with (format csv);
481,click
491,view
502,click
51\.
52```
53 
54Reference: [COPY](https://www.postgresql.org/docs/current/sql-copy.html)
55