Tuples

CQL tuples are ordered sets of anonymous, typed fields. They can be used as a column type in tables, or a field type in user-defined types:

CREATE TABLE ks.collect_things (
  pk int,
  ck1 text,
  ck2 text,
  v tuple<int, text, float>,
  PRIMARY KEY (pk, ck1, ck2)
);

Fetching tuples from results

The driver maps tuple columns to the TupleValue class, which exposes getters and setters to access individual fields by index:

Row row = session.execute("SELECT v FROM ks.collect_things WHERE pk = 1").one();

TupleValue tupleValue = row.getTupleValue("v");
int field0 = tupleValue.getInt(0);
String field1 = tupleValue.getString(1);
Float field2 = tupleValue.getFloat(2);

Using tuples as parameters

Statements may contain tuples as bound values:

PreparedStatement ps =
  session.prepare(
      "INSERT INTO ks.collect_things (pk, ck1, ck2, v) VALUES (:pk, :ck1, :ck2, :v)");

To create a new tuple value, you must first have a reference to its TupleType. There are various ways to get it:

• from the statement’s metadata

  TupleType tupleType = (TupleType) ps.getVariableDefinitions().get("v").getType();
  

• from the driver’s schema metadata:

  TupleType tupleType =
      (TupleType)
          session
              .getMetadata()
              .getKeyspace("ks")
              .getTable("collect_things")
              .getColumn("v")
              .getType();
  

• from another tuple value:

  TupleType tupleType = tupleValue.getType();
  

• or creating it from scratch:

  TupleType tupleType = DataTypes.tupleOf(DataTypes.INT, DataTypes.TEXT, DataTypes.FLOAT);
  

  Note that the resulting type is detached.

Once you have the type, call newValue() and set the fields:

TupleValue tupleValue =
    tupleType.newValue().setInt(0, 1).setString(1, "hello").setFloat(2, 2.3f);

// Or as a one-liner for convenience:
TupleValue tupleValue = tupleType.newValue(1, "hello", 2.3f);

And bind your tuple value like any other type:

BoundStatement bs =
    ps.boundStatementBuilder()
        .setInt("pk", 1)
        .setString("ck1", "1")
        .setString("ck2", "1")
        .setTupleValue("v", tupleValue)
        .build();
session.execute(bs);

Tuples are also used for multi-column IN restrictions (usually for tables with composite clustering keys):

PreparedStatement ps =
    session.prepare("SELECT * FROM ks.collect_things WHERE pk = 1 and (ck1, ck2) IN (:choice1, :choice2)");

TupleType tupleType = DataTypes.tupleOf(DataTypes.TEXT, DataTypes.TEXT);
BoundStatement bs = ps.boundStatementBuilder()
    .setTupleValue("choice1", tupleType.newValue("a", "b"))
    .setTupleValue("choice2", tupleType.newValue("c", "d"))
    .build();

If you bind the whole list of choices as a single variable, a list of tuple values is expected:

PreparedStatement ps =
    // Note the absence of parentheses around ':choices'
    session.prepare("SELECT * FROM ks.collect_things WHERE pk = 1 and (ck1, ck2) IN :choices");

TupleType tupleType = DataTypes.tupleOf(DataTypes.TEXT, DataTypes.TEXT);
List<TupleValue> choices = new ArrayList<>();
choices.add(tupleType.newValue("a", "b"));
choices.add(tupleType.newValue("c", "d"));
BoundStatement bs =
    ps.boundStatementBuilder().setList("choices", choices, TupleValue.class).build();