ABORT

Synopsis

Use the ABORT statement to roll back the current transaction and discards all updates by the transaction.

abort ::= ABORT [ TRANSACTION | WORK ]

Add optional keyword — has no effect.

Add optional keyword — has no effect.

Create a sample table.

yugabyte=# CREATE TABLE sample(k1 int, k2 int, v1 int, v2 text, PRIMARY KEY (k1, k2));

Begin a transaction and insert some rows.

yugabyte=# BEGIN TRANSACTION; SET TRANSACTION ISOLATION LEVEL REPEATABLE READ;

yugabyte=# INSERT INTO sample(k1, k2, v1, v2) VALUES (1, 2.0, 3, 'a'), (1, 3.0, 4, 'b');

Open the YSQL shell (ysqlsh) and begin another transaction to insert some more rows.

yugabyte=# BEGIN TRANSACTION; SET TRANSACTION ISOLATION LEVEL REPEATABLE READ;

yugabyte=# INSERT INTO sample(k1, k2, v1, v2) VALUES (2, 2.0, 3, 'a'), (2, 3.0, 4, 'b');

In each shell, check the only the rows from the current transaction are visible.

First shell.

yugabyte=# SELECT * FROM sample; -- run in first shell

 k1 | k2 | v1 | v2
----+----+----+----
  1 |  2 |  3 | a
  1 |  3 |  4 | b
(2 rows)

Second shell.

yugabyte=# SELECT * FROM sample; -- run in second shell

 k1 | k2 | v1 | v2
----+----+----+----
  2 |  2 |  3 | a
  2 |  3 |  4 | b
(2 rows)

Commit the first transaction and abort the second one.

yugabyte=# COMMIT TRANSACTION; -- run in first shell.

Abort the current transaction (from the first shell).

yugabyte=# ABORT TRANSACTION; -- run second shell.

In each shell, check that only the rows from the committed transaction are visible.

yugabyte=# SELECT * FROM sample; -- run in first shell.

 k1 | k2 | v1 | v2
----+----+----+----
  1 |  2 |  3 | a
  1 |  3 |  4 | b
(2 rows)

yugabyte=# SELECT * FROM sample; -- run in second shell.

 k1 | k2 | v1 | v2
----+----+----+----
  1 |  2 |  3 | a
  1 |  3 |  4 | b
(2 rows)