Thanks, this client library is really great! It should be included in the Apache Cassandra Wiki's "High level clients" page (http://wiki.apache.org/cassandra/ClientOptions).
Anyway, here is the code I created (using the cassandra-driver-core Maven artifact).

Cluster cluster = Cluster.builder().addContactPoints("127.0.0.1").build();
Session session = cluster.connect("graph");
for (Row row : session.execute("SELECT * FROM vertices")) {
System.out.println("id: " + row.getString("id"));
System.out.println("properties: ");
Map<String, String> map = row.getMap("properties", String.class, String.class);
for (Map.Entry<String, String> entry : map.entrySet()) {
System.out.println("- " + entry.getKey() + ": " + entry.getValue());
}
}

Gabor


On 25 March 2013 14:52, Sylvain Lebresne <sylvain@datastax.com> wrote:
I would advise you not to use raw thrift. It's just a low-level transport as far as CQL3 is concerned, and what you will get is binary encoded data that you will have to decode manually. Use a client library (like https://github.com/datastax/java-driver) that will do that for you.

Though to answer your initial question, the binary format used by map (that you will have to decode yourself) is described in https://git-wip-us.apache.org/repos/asf?p=cassandra.git;a=blob;f=doc/native_protocol.spec (Section 6).

--
Sylvain


On Mon, Mar 25, 2013 at 2:30 PM, Szárnyas Gábor <szarnyasg@gmail.com> wrote:
Hello!

I got stuck when trying to query CQL3 collections from Java. 
I'm using Cassandra 1.2.3 with CQL3. I created a column family to store a property graph's edges with the following command:

CREATE TABLE vertices ( 
  id text PRIMARY KEY, 
  properties map<text, text>
)

I can access the data from the cqlsh, however, I couldn't figure out how to iterate through the map entries in Java. 
The following code iterates through the rows and columns, but does not retrieve the key-value pairs of the "properties" map.

String query = "SELECT * FROM vertices";
CqlResult cqlResult = client.execute_cql3_query(ByteBuffer.wrap(query.getBytes()), COMPRESSION_LEVEL.NONE, CONSISTENCY_LEVEL.ALL);

Iterator<CqlRow> rowsIterator = cqlResult.getRowsIterator();
while (rowsIterator.hasNext()) {
  CqlRow cqlRow = rowsIterator.next();
  Iterator<Column> columnsIterator = cqlRow.getColumnsIterator();
  while (columnsIterator.hasNext()) {
    Column cqlColumn = columnsIterator.next();

    byte[] name = cqlColumn.getName();
    String nameString = new String(name);
    System.out.print(nameString + ": ");

    byte[] value = cqlColumn.getValue();
    String string = new String(value);
    System.out.println(string);
  }
}

The cqlResult.getSchema() method shows the column with the type "org.apache.cassandra.db.marshal.MapType(org.apache.cassandra.db.marshal.UTF8Type,org.apache.cassandra.db.marshal.UTF8Type)". How can I create a HashMap<String, String> from each row's properties cell?

Thanks,
Gabor