INSERT VALUES statement in ksqlDB for Confluent Platform¶

Synopsis¶

INSERT INTO <stream_name|table_name> [(column_name [, ...]])]
  VALUES (value [,...]);

Description¶

Produce a row into an existing stream or table and its underlying topic based on explicitly specified values.

If column names are specified, the order of the values must match the order of the names. Any column not explicitly given a value is set to null. If a column is given a null value or is not set, it does not act as a tombstone.

Tombstones are not supported with INSERT INTO. Pseudo columns are supported on a case by case basis. ROWTIME, may be provided. However, ROWPARTITION and ROWOFFSET are disallowed. Header columns are also disallowed.

If no columns are specified, a value for every column is expected in the same order as the schema, with key columns first.

Note

ROWTIME may be specified as an explicit column but isn’t required when you omit the column specifications. If not supplied, it defaults to the local machine time.

Example¶

The following statements are valid for a source with a schema like KEY_COL VARCHAR KEY, COL_A VARCHAR.

-- inserts (ROWTIME:=1510923225000, KEY_COL:="key", COL_A:="A")
INSERT INTO foo (ROWTIME, KEY_COL, COL_A) VALUES (1510923225000, 'key', 'A');

-- also inserts (ROWTIME:=1510923225000, KEY_COL:="key", COL_A:="A")
INSERT INTO foo (COL_A, ROWTIME, KEY_COL) VALUES ('A', 1510923225000, 'key');

-- inserts (ROWTIME:=current_time(), KEY_COL:="key", COL_A:="A")
INSERT INTO foo VALUES ('key', 'A');

-- inserts (ROWTIME:=current_time(), KEY_COL:="key", COL_A:="A")
INSERT INTO foo (KEY_COL, COL_A) VALUES ('key', 'A');

-- inserts (ROWTIME:=current_time(), KEY_COL:="key", COL_A:=null)
INSERT INTO foo (KEY_COL) VALUES ('key');

The values are serialized by using the format(s) specified in the original CREATE statement.

Tip

See INSERT INTO VALUES in action: