me.prettyprint.cassandra.model

Class CqlQuery<K,N,V>

java.lang.Object

me.prettyprint.cassandra.model.AbstractBasicQuery<K,N,CqlRows<K,N,V>>

me.prettyprint.cassandra.model.CqlQuery<K,N,V>

All Implemented Interfaces:

Query<CqlRows<K,N,V>>

public class CqlQuery<K,N,V>
extends AbstractBasicQuery<K,N,CqlRows<K,N,V>>

First cut at a CQL implementation. Not too much time has been spent here as this API is currently a moving target. We have a lot of experience with these hijinks by the Apache Cassandra team, so it was deemed prudent to do something simple initially until the dust settles. You are expected to know what you are getting into if you plan on using CQL queries in your application. Spend some time looking through the unit tests here in Hector and the Cassandra source tree. For a number of detailed examples, see test_cql.py in the test/system folder of the Apache Cassandra source distribution. Note: if you immediately get an exception such as: "InvalidRequestException(why:cannot parse 'foo' as hex bytes)" It means one of two things:

1. you have not formatted your query correct
2. You have not configured the correct validators on your column family

In both cases, even though the query is most likely a string, it is up to you to format this query according to the comparator (used for the column name), key validator and value validator. This can be a little confusing as only the comparator is required. The other two default to BytesType. See the docs on CqlRows for additional details.

Author:

zznate

Field Summary

Fields inherited from class me.prettyprint.cassandra.model.AbstractBasicQuery

columnFamilyName, columnNameSerializer, consistency, cql3, cqlVersion, keySerializer, keyspace

Constructor Summary

Constructors

Constructor and Description
CqlQuery(Keyspace k, Serializer<K> keySerializer, Serializer<N> nameSerializer, Serializer<V> valueSerializer)

Method Summary

Methods

Modifier and Type	Method and Description
QueryResult<CqlRows<K,N,V>>	execute() For future releases, we should consider refactoring how execute method converts the thrift CqlResult to Hector CqlRows.
protected HConsistencyLevel	getConsistency()
CqlQuery<K,N,V>	setCqlVersion(String version) Method will check to see if the version starts with "3" and set the boolean cql3 to true.
CqlQuery<K,N,V>	setQuery(ByteBuffer qeury)
CqlQuery<K,N,V>	setQuery(String query) Set the query as a String.
CqlQuery<K,N,V>	setSuppressKeyInColumns(boolean suppressKeyInColumns)
CqlQuery<K,N,V>	useCompression()

Methods inherited from class me.prettyprint.cassandra.model.AbstractBasicQuery

getColumnNameSerializer, getConsistencyLevel, getCqlVersion, getKeySerializer, setColumnFamily, setColumnNameSerializer, setConsistencyLevel, setKeySerializer

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

CqlQuery

public CqlQuery(Keyspace k,
        Serializer<K> keySerializer,
        Serializer<N> nameSerializer,
        Serializer<V> valueSerializer)

Method Detail

setQuery

public CqlQuery<K,N,V> setQuery(String query)

Set the query as a String. Here for convienience. See above for some caveats. Calls StringSerializer.toByteBuffer(String) directly.

Parameters:

query -

Returns:

setQuery

public CqlQuery<K,N,V> setQuery(ByteBuffer qeury)

setCqlVersion

public CqlQuery<K,N,V> setCqlVersion(String version)

Method will check to see if the version starts with "3" and set the boolean cql3 to true. This makes it easier for execute method and avoids doing lots of cqlVersion.startsWith() tests.

Overrides:

setCqlVersion in class AbstractBasicQuery<K,N,CqlRows<K,N,V>>

setSuppressKeyInColumns

public CqlQuery<K,N,V> setSuppressKeyInColumns(boolean suppressKeyInColumns)

useCompression

public CqlQuery<K,N,V> useCompression()

getConsistency

protected HConsistencyLevel getConsistency()

execute

public QueryResult<CqlRows<K,N,V>> execute()

For future releases, we should consider refactoring how execute method converts the thrift CqlResult to Hector CqlRows. Starting with CQL3 the KEY is no longer returned by default and is no longer treated as a "special" thing. To get the KEY from the resultset, we have to look at the columns. Using the KEY in CqlRows is nice, and makes it easy to get results by the KEY. The downside is that users have to explicitly include the KEY in the statement. Changing how CqlRows work "might" break some existing use cases. For now, the implementation finds the column and expects the statement to have the KEY.