Cassandra - Revision history

Admin: /* Scaling */

2017-09-15T20:14:21Z

Scaling

← Older revision		Revision as of 20:14, 15 September 2017
Line 97:		Line 97:

	Cassandra uses a similar architecture to Hadoop, in that it distributes data over multiple nodes, and uses replication to safeguard against data loss due to the failure of nodes in the cluster. A single row that is written to the database is the responsibility of a single node (row-to-node mapping is determined by your partitioning scheme, see info about "snitch" above). However, that piece of data is mirrored to replica nodes in the cluster. The user can configure the replication factor, but by default it is 3. If/when the primary node fails, the data that was stored on it can still be accessed via a replica node.		Cassandra uses a similar architecture to Hadoop, in that it distributes data over multiple nodes, and uses replication to safeguard against data loss due to the failure of nodes in the cluster. A single row that is written to the database is the responsibility of a single node (row-to-node mapping is determined by your partitioning scheme, see info about "snitch" above). However, that piece of data is mirrored to replica nodes in the cluster. The user can configure the replication factor, but by default it is 3. If/when the primary node fails, the data that was stored on it can still be accessed via a replica node.

			==Snitches==

			Cassandra configures data center, rack, and node layouts using "snitches".

			Link: http://docs.datastax.com/en/cassandra/2.1/cassandra/architecture/architectureSnitchesAbout_c.html

	=Basic Usage=		=Basic Usage=

Admin: /* CQL Examples */

2017-09-15T19:53:40Z

CQL Examples

← Older revision		Revision as of 19:53, 15 September 2017
Line 77:		Line 77:
	CREATE KEYSPACE mydb WITH replication =		CREATE KEYSPACE mydb WITH replication =
	{'class': 'NetworkTopologyStrategy', 'data_center_1' : 3, 'data_center_2' : 3};		{'class': 'NetworkTopologyStrategy', 'data_center_1' : 3, 'data_center_2' : 3};
			</pre>

			Note that the names of the data centers are configured when you create the snitch. To print out the data center names and rack numbers of each node, use the command:

			<pre>
			nodetool status
	</pre>		</pre>

Admin: /* CQL Examples */

2017-09-15T19:52:45Z

CQL Examples

← Older revision		Revision as of 19:52, 15 September 2017
Line 74:		Line 74:
	{'class': 'SimpleStrategy', 'replication_factor': 1};		{'class': 'SimpleStrategy', 'replication_factor': 1};

	/* Cassandra - Create new keyspace with ~~simple strategy, not replicating any~~ data */		/* Cassandra - Create new keyspace with 3x replication in each data center */
	CREATE KEYSPACE mydb WITH replication =		CREATE KEYSPACE mydb WITH replication =
	{'class': 'NetworkTopologyStrategy', 'data_center_1' : 3, 'data_center_2' : 3};		{'class': 'NetworkTopologyStrategy', 'data_center_1' : 3, 'data_center_2' : 3};

Admin: /* CQL Examples */

2017-09-15T19:52:14Z

CQL Examples

← Older revision		Revision as of 19:52, 15 September 2017
Line 67:		Line 67:
	However, CQL uses Cassandra's concept of a keyspace - analogous to an SQL database - and Cassandra's concept of a column family - analogous to an SQL table.		However, CQL uses Cassandra's concept of a keyspace - analogous to an SQL database - and Cassandra's concept of a column family - analogous to an SQL table.

	~~Compare~~ CQL ~~with SQL~~: ~~CQL requires us to specify how Cassandra should handle the replication of data.~~		Here's how we create a keyspace in Cassandra using CQL:

	<pre>		<pre>
	/* Cassandra - Create new keyspace */		/* Cassandra - Create new keyspace with simple strategy, not replicating any data */
	CREATE KEYSPACE mydb WITH replication =		CREATE KEYSPACE mydb WITH replication =
	{'class': 'SimpleStrategy', 'replication_factor': 1};		{'class': 'SimpleStrategy', 'replication_factor': 1};

			/* Cassandra - Create new keyspace with simple strategy, not replicating any data */
			CREATE KEYSPACE mydb WITH replication =
			{'class': 'NetworkTopologyStrategy', 'data_center_1' : 3, 'data_center_2' : 3};
			</pre>

			(Link to details of keyspace creation: [http://docs.datastax.com/en/cql/3.1/cql/cql_reference/create_keyspace_r.html])

			Compare this to the way we create a new database in SQL, which is simpler:

			<pre>
	/* SQL - Create new database */		/* SQL - Create new database */
	CREATE DATABASE mydb;		CREATE DATABASE mydb;

Admin: /* CQL Examples */

2017-09-15T19:49:04Z

CQL Examples

← Older revision		Revision as of 19:49, 15 September 2017
Line 67:		Line 67:
	However, CQL uses Cassandra's concept of a keyspace - analogous to an SQL database - and Cassandra's concept of a column family - analogous to an SQL table.		However, CQL uses Cassandra's concept of a keyspace - analogous to an SQL database - and Cassandra's concept of a column family - analogous to an SQL table.

	~~To create a database~~:		Compare CQL with SQL: CQL requires us to specify how Cassandra should handle the replication of data.

	<pre>		<pre>

Admin: /* CQL Examples */

2017-09-15T19:48:29Z

CQL Examples

@@ Line 63: / Line 63: @@
 /* Delete Data */
 DELETE FROM someTable WHERE someField = 100;
 </pre>

Admin: /* CQL Examples */

2017-09-15T19:46:30Z

CQL Examples

← Older revision		Revision as of 19:46, 15 September 2017
Line 47:		Line 47:

	/* Alter Tables /*		/* Alter Tables /*
	ALTER TABLE someTable ADD ~~myField~~ INT;		ALTER TABLE someTable ADD someField INT;

	/* Create Indexes */		/* Create Indexes */

Admin: /* CQL */

2017-09-15T19:46:12Z

CQL

@@ Line 35: / Line 35: @@
 * No NOT operators
 * The AND operator is the only operator available
 ==Scaling==

Admin: /* Resources */

2017-09-15T19:43:37Z

Resources

← Older revision		Revision as of 19:43, 15 September 2017
Line 108:		Line 108:
	Nice blog writeup on results of Cassandra benchmarking: http://blog.markedup.com/2013/02/cassandra-hive-and-hadoop-how-we-picked-our-analytics-stack/		Nice blog writeup on results of Cassandra benchmarking: http://blog.markedup.com/2013/02/cassandra-hive-and-hadoop-how-we-picked-our-analytics-stack/

			Medium essay on CQL: https://medium.com/@alexbmeng/cassandra-query-language-cql-vs-sql-7f6ed7706b4c

Admin: /* About Cassandra */

2017-09-15T18:49:52Z

About Cassandra

← Older revision		Revision as of 18:49, 15 September 2017
Line 5:		Line 5:
	=About Cassandra=		=About Cassandra=

	Cassandra is a scalable database tool. It is designed for ''extremely'' large data sets.		Cassandra is a scalable database tool. It is designed to achieve linear scaling for ''extremely'' large data sets.

	==NoSQL==		==NoSQL==
Line 36:		Line 36:
	* The AND operator is the only operator available		* The AND operator is the only operator available

	==~~=Types=~~==		==Scaling==

	Cassandra ~~supports several~~ data ~~types:~~		Cassandra uses a similar architecture to Hadoop, in that it distributes data over multiple nodes, and uses replication to safeguard against data loss due to the failure of nodes in the cluster. A single row that is written to the database is the responsibility of a single node (row-to-node mapping is determined by your partitioning scheme, see info about "snitch" above). However, that piece of data is mirrored to replica nodes in the cluster. The user can configure the replication factor, but by default it is 3. If/when the primary node fails, the data that was stored on it can still be accessed via a replica node.
	* native types
	* collection types
	* user~~-defined types~~
	* tuple types
	* custom types

	=Basic Usage=		=Basic Usage=