Query methods

Annotate a DAO method with @Query to provide your own query string:

public interface SensorReadingDao {
  @Query("SELECT count(*) FROM sensor_readings WHERE id = :id")
  long countById(int id);

This is the equivalent of what was called “accessor methods” in the driver 3 mapper.


The query string will typically contain CQL placeholders. The method’s parameters must match those placeholders: same name and a compatible Java type.

@Query("SELECT count(*) FROM sensor_readings WHERE id = :id AND year = :year")
long countByIdAndYear(int id, int year);

The annotation can define a null saving strategy that applies to the method parameters.

A Function<BoundStatementBuilder, BoundStatementBuilder> or UnaryOperator<BoundStatementBuilder> can be added as the last parameter. It will be applied to the statement before execution. This allows you to customize certain aspects of the request (page size, timeout, etc) at runtime. See statement attributes.

Return type

The method can return:

  • void.

  • a boolean or Boolean, which will be mapped to ResultSet#wasApplied(). This is intended for conditional queries.

  • a long or Long, which will be mapped to the first column of the first row, expecting CQL type BIGINT. This is intended for count queries. The method will fail if the result set is empty, or does not match the expected format.

  • a Row. This means the result is not converted, the mapper only extracts the first row of the result set and returns it. The method will return null if the result set is empty.

  • a single instance of an Entity class. The method will extract the first row and convert it, or return null if the result set is empty.

  • an Optional of an entity class. The method will extract the first row and convert it, or return Optional.empty() if the result set is empty.

  • a ResultSet. The method will return the raw query result, without any conversion.

  • a BoundStatement. This is intended for queries where you will execute this statement later or in a batch.

  • a PagingIterable. The method will convert each row into an entity instance.

  • a Stream. The method will convert each row into an entity instance. For details about the stream’s characteristics, see PagingIterable.spliterator.

  • a CompletionStage or CompletableFuture of any of the above. The method will execute the query asynchronously. Note that for result sets and iterables, you need to switch to the asynchronous equivalent AsyncResultSet and MappedAsyncPagingIterable respectively.

  • a ReactiveResultSet, or a MappedReactiveResultSet of the entity class.

  • a custom type.

Target keyspace and table

To avoid hard-coding the keyspace and table name, the query string supports 3 additional placeholders: ${keyspaceId}, ${tableId} and ${qualifiedTableId}. They get substituted at DAO initialization time, with the keyspace and table that the DAO was built with.

For example, given the following:

public interface TestDao {
  @Query("SELECT * FROM ${keyspaceId}.${tableId}")
  ResultSet queryFromKeyspaceAndTable();

  @Query("SELECT * FROM ${qualifiedTableId}")
  ResultSet queryFromQualifiedTable();

public interface TestMapper {
  TestDao dao(@DaoKeyspace String keyspace, @DaoTable String table);

  TestDao dao(@DaoTable String table);

TestDao dao1 = mapper.dao("ks", "t");
TestDao dao2 = mapper.dao("t");


  • dao1.queryFromKeyspaceAndTable() and dao1.queryFromQualifiedTable() both execute SELECT * FROM ks.t.
  • dao2.queryFromKeyspaceAndTable() fails: no keyspace was specified for this DAO, so ${keyspaceId} can’t be substituted.
  • dao1.queryFromQualifiedTable() executes SELECT * FROM t. In other words, ${qualifiedTableId} uses the keyspace if it is available, but resolves to the table name only if it isn’t. Whether the query succeeds or not depends on whether the session that the mapper was built with has a default keyspace.