To start an UPDATE query, use one of the update methods in QueryBuilder. There are several variants depending on whether your table name is qualified, and whether you use identifiers or raw strings:

import static com.datastax.oss.driver.api.querybuilder.QueryBuilder.*;

UpdateStart update = update("user");

Note that, at this stage, the query can’t be built yet. You need at least one assignment and one relation.


The USING TIMESTAMP clause comes right after the table, and specifies the timestamp at which the mutation will be applied. You can pass either a literal value:


Or a bind marker:


If you call the method multiple times, the last value will be used.

Time To Live (TTL)

You can generate a USING TTL clause that will cause column values to be deleted (marked with a tombstone) after the specified time (in seconds) has expired. This can be done with a literal:

update("user").usingTtl(60).setColumn("v", bindMarker()).whereColumn("k").isEqualTo(bindMarker());
// UPDATE user USING TTL 60 SET v=? WHERE k=?

Or a bind marker:

update("user").usingTtl(bindMarker()).setColumn("v", bindMarker()).whereColumn("k").isEqualTo(bindMarker());
// UPDATE user USING TTL ? SET v=? WHERE k=?

You can clear a previously set TTL by setting the value to 0:

update("user").usingTtl(0).setColumn("v", bindMarker()).whereColumn("k").isEqualTo(bindMarker());
// UPDATE user USING TTL 0 SET v=? WHERE k=?

Setting the value to 0 will result in removing the TTL from the column in Cassandra when the query is executed. This is distinctly different than setting the value to null. Passing a null value to this method will only remove the USING TTL clause from the query, which will not alter the TTL (if one is set) in Cassandra.


An assignment is an operation that appears after the SET keyword. You need at least one for a valid update query.

The easiest way to add an assignment is with one of the fluent methods:

update("user").setColumn("v", bindMarker());
// UPDATE user SET v=?...

You can also create it manually with one of the factory methods in Assignment, and then pass it to set():

    Assignment.setColumn("v", bindMarker()));
// UPDATE user SET v=?...

If you have multiple assignments, you can add them all in a single call. This is a bit more efficient since it creates less temporary objects:

    Assignment.setColumn("v1", bindMarker()),
    Assignment.setColumn("v2", bindMarker()))
// UPDATE user SET v1=?,v2=?...

Simple columns

As shown already, setColumn changes the value of a column. It can take a bind marker or a literal (which must have the same CQL type as the column):

update("user").setColumn("last_name", literal("Doe"));
// UPDATE user SET last_name='Doe'...

UDT fields

setField modifies a field inside of a UDT column:

update("user").setField("address", "street", bindMarker());
// UPDATE user SET address.street=?...


Counter columns can be incremented by a given amount:

update("foo").increment("c", bindMarker());
// UPDATE foo SET c+=?...

update("foo").increment("c", literal(4));
// UPDATE foo SET c+=4...

There is a shortcut to increment by 1:

// UPDATE foo SET c+=1...

All those methods have a decrement counterpart:

// UPDATE foo SET c-=1...


mapValue changes a value in a map. The key is expressed as a term (here a literal value) :

update("product").setMapValue("features", literal("color"), bindMarker())
// UPDATE product SET features['color']=?...

append operates on any CQL collection type (list, set or map). If you pass a literal, it must also be a collection, with the same CQL type of elements:

update("foo").append("l", bindMarker());
// UPDATE foo SET l+=?...

List<Integer> value = Arrays.asList(1, 2, 3);
update("foo").append("l", literal(value));
// UPDATE foo SET l+=[1,2,3]...

Set<Integer> value = new HashSet<>(Arrays.asList(1, 2, 3));
update("foo").append("s", literal(value))
// UPDATE foo SET s+={1,2,3}...

Map<Integer, String> value = new HashMap<>();
value.put(1, "bar");
value.put(2, "baz");
update("foo").append("m", literal(value));
// UPDATE foo SET m+={1:'bar',2:'baz'}...

If you only have one element to append, there are shortcuts to avoid creating a collection in your code:

update("foo").appendListElement("l", literal(1));
// UPDATE foo SET l+=[1]...

update("foo").appendSetElement("s", literal(1));
// UPDATE foo SET s+={1}...

update("foo").appendMapEntry("m", literal(1), literal("bar"));
// UPDATE foo SET m+={1:'bar'}...

All those methods have a prepend counterpart:

update("foo").prepend("l", bindMarker());
// UPDATE foo SET l=?+l...

As well as a remove counterpart:

update("foo").remove("l", bindMarker());
// UPDATE foo SET l-=?...


Once you have at least one assignment, relations can be added with the fluent whereXxx() methods:

update("foo").setColumn("v", bindMarker()).whereColumn("k").isEqualTo(bindMarker());
// UPDATE foo SET v=? WHERE k=?

Or you can build and add them manually:

update("foo").setColumn("v", bindMarker()).where(
// UPDATE foo SET v=? WHERE k=?

Once there is at least one assignment and one relation, the statement can be built:

SimpleStatement statement = update("foo")
    .setColumn("k", bindMarker())

Relations are a common feature used by many types of statements, so they have a dedicated page in this manual.


Conditions get added with the fluent ifXxx() methods:

    .setColumn("v", bindMarker())
// UPDATE foo SET v=? WHERE k=? IF v=?

Or you can build and add them manually:

    .setColumn("v", bindMarker())
// UPDATE foo SET v=? WHERE k=? IF v=?

Conditions are a common feature used by UPDATE and DELETE, so they have a dedicated page in this manual.