UPSERT

On this page

The UPSERT statement is short-hand for INSERT ON CONFLICT. It inserts rows in cases where specified values do not violate uniqueness constraints, and it updates rows in cases where values do violate uniqueness constraints.

Considerations

• UPSERT considers uniqueness only for Primary Key columns. INSERT ON CONFLICT is more flexible and can be used to consider uniqueness for other columns. For more details, see How UPSERT Transforms into INSERT ON CONFLICT below.

• When inserting/updating all columns of a table, and the table has no secondary indexes, UPSERT will be faster than the equivalent INSERT ON CONFLICT statement, as it will write without first reading. This may be particularly useful if you are using a simple SQL table of two columns to simulate direct KV access.

Required Privileges

The user must have the INSERT and UPDATE privileges on the table.

Synopsis

Parameters

How UPSERT Transforms into INSERT ON CONFLICT

UPSERT considers uniqueness only for primary key columns. For example, assuming that columns a and b are the primary key, the following UPSERT and INSERT ON CONFLICT statements are equivalent:

> UPSERT INTO t (a, b, c) VALUES (1, 2, 3);

> INSERT INTO t (a, b, c)
    VALUES (1, 2, 3)
    ON CONFLICT (a, b)
    DO UPDATE SET c = excluded.c;

INSERT ON CONFLICT is more flexible and can be used to consider uniqueness for columns not in the primary key. For more details, see the Upsert that Fails (Conflict on Non-Primary Key) example below.

Examples

Upsert that Inserts a Row (No Conflict)

In this example, the id column is the primary key. Because the inserted id value does not conflict with the id value of any existing row, the UPSERT statement inserts a new row into the table.

> SELECT * FROM accounts;

+----+----------+
| id | balance  |
+----+----------+
|  1 |  10000.5 |
|  2 | 20000.75 |
+----+----------+

> UPSERT INTO accounts (id, balance) VALUES (3, 6325.20);

> SELECT * FROM accounts;

+----+----------+
| id | balance  |
+----+----------+
|  1 |  10000.5 |
|  2 | 20000.75 |
|  3 |   6325.2 |
+----+----------+

Upsert that Updates a Row (Conflict on Primary Key)

In this example, the id column is the primary key. Because the inserted id value is not unique, the UPSERT statement updates the row with the new balance.

> SELECT * FROM accounts;

+----+----------+
| id | balance  |
+----+----------+
|  1 |  10000.5 |
|  2 | 20000.75 |
|  3 |   6325.2 |
+----+----------+

> UPSERT INTO accounts (id, balance) VALUES (3, 7500.83);

> SELECT * FROM accounts;

+----+----------+
| id | balance  |
+----+----------+
|  1 |  10000.5 |
|  2 | 20000.75 |
|  3 |  7500.83 |
+----+----------+

Upsert that Fails (Conflict on Non-Primary Key)

UPSERT will not update rows when the uniquness conflict is on columns not in the primary key. In this example, the a column is the primary key, but the b column also has the Unique constraint. Because the inserted b value is not unique, the UPSERT fails.

> SELECT * FROM unique_test;

+---+---+
| a | b |
+---+---+
| 1 | 1 |
| 2 | 2 |
| 3 | 3 |
+---+---+

> UPSERT INTO unique_test VALUES (4, 1);

pq: duplicate key value (b)=(1) violates unique constraint "unique_test_b_key"

In such a case, you would need to use the INSERT ON CONFLICT statement to specify the b column as the column with the Unique constraint.

> INSERT INTO unique_test VALUES (4, 1) ON CONFLICT (b) DO UPDATE SET a = excluded.a;

> SELECT * FROM unique_test;

+---+---+
| a | b |
+---+---+
| 2 | 2 |
| 3 | 3 |
| 4 | 1 |
+---+---+

See Also

• INSERT
• Other SQL Statements

Was this helpful?