Skip to main content
Version: 6.0-SNAPSHOT

Using the generic MapStore

With the generic MapStore, you can configure a map to cache data from and write data back to an external system. This topic includes an example of how to configure a map with a generic MapStore that connects to a MySQL database.

note

The objects created in the distributed map are stored as GenericRecord. You can use the type-name property to store the data in a POJO (Plain Old Java Object).

important

GenericMapStore is designed to handle GenericRecord and does not support automatic transformation between stored data (GenericRecord) and domain objects (POJOs) in embedded mode. While the type-name property allows specifying a Compact type name, it does not enable domain object mapping.

For a list of all supported external systems, including databases, see available data connection types.

Before you begin​

You need a data connection that's configured on all cluster members.

Add dependencies​

If you are using a Hazelcast JAR file, you need to ensure the following is added to your classpath:

xml
<dependency>
<groupId>com.hazelcast</groupId>
<artifactId>hazelcast-sql</artifactId>
</dependency>

<dependency>
<groupId>com.hazelcast</groupId>
<artifactId>hazelcast-mapstore</artifactId>
</dependency>
note

If you are using the slim distribution, you need to add hazelcast-mapstore. If you are using MongoDb, you also need to add hazelcast-jet-mongodb.

Quickstart configuration​

This example shows a basic map configuration that uses a data connection called my-mysql-database. See Distributed Map for the details of other properties that you include in your map configuration.

xml
<hazelcast>
...
<map name="myMapName">
<map-store enabled="true">
<class-name>com.hazelcast.mapstore.GenericMapStore</class-name> (1)
<properties>
<property name="data-connection-ref">my-mysql-database</property> (2)
</properties>
</map-store>
</map>
...
</hazelcast>
1 The class name of the generic MapStore that's built into Hazelcast. This implementation reads from the data connection configuration to create a SQL mapping.
2 The name of your data connection.

SQL mapping for the generic MapStore​

When you configure a map with the generic MapStore, Hazelcast creates a SQL mapping with the JDBC connector. The name of the mapping is the same name as your map prefixed with __map-store.. This mapping is used to read data from or write data to the external system and is removed whenever the configured map is removed. You can also configure this SQL mapping, using configuration properties.

Configuration properties for the generic MapStore​

These configuration properties allow you to configure the generic MapStore and its SQL mapping.

tip

As well as these properties, you can also configure general MapStore settings.

note

Property names are case-sensitive.

OptionDescriptionDefaultExample
data-connection-ref (required)The name of the data connection that sets up a mapping. Use this property when you have an existing data connection to an external system.'' (empty)
xml
<hazelcast>
<map name="myMapName">
<map-store enabled="true">
<class-name>com.hazelcast.mapstore.GenericMapStore</class-name>
<properties>
<property name="data-connection-ref">my-mysql-database</property>
</properties>
</map-store>
</hazelcast>
external-nameExternal name of the data (e.g. table or collection) to read from. May be case sensitive, depending on database used.The name of the map.
xml
<hazelcast>
<map name="myMapName">
<map-store enabled="true">
<class-name>com.hazelcast.mapstore.GenericMapStore</class-name>
<properties>
<property name="data-connection-ref">my-mysql-database</property>
<property name="external-name">test</property>
</properties>
</map-store>
</hazelcast>
mapping-typeSQL connector to use for the mapping.The SQL connector is derived from the data connection in the configuration.
xml
<hazelcast>
<map name="myMapName">
<map-store enabled="true">
<class-name>com.hazelcast.mapstore.GenericMapStore</class-name>
<properties>
<property name="data-connection-ref">my-mysql-database</property>
<property name="mapping-type">JDBC</property>
</properties>
</map-store>
</hazelcast>
id-columnName of the column that contains the primary key. Do not specify a column without a unique key constraint as the id-column. May be case sensitive, depending on database used.id
xml
<hazelcast>
<map name="myMapName">
<map-store enabled="true">
<class-name>com.hazelcast.mapstore.GenericMapStore</class-name>
<properties>
<property name="data-connection-ref">my-mysql-database</property>
<property name="id-column">id</property>
</properties>
</map-store>
</hazelcast>
columnsNames of the columns to map. This value must include a subset of columns in the table. Missing columns must have a default value defined. May be case sensitive, depending on database used.
xml
<hazelcast>
<map name="myMapName">
<map-store enabled="true">
<class-name>com.hazelcast.mapstore.GenericMapStore</class-name>
<properties>
<property name="data-connection-ref">my-mysql-database</property>
<property name="columns">name</property>
</properties>
</map-store>
</hazelcast>
type-nameThe type name of the compact GenericRecord. Use this property to map your record to an existing domain class.The name of the map.
xml
<hazelcast>
<map name="myMapName">
<map-store enabled="true">
<class-name>com.hazelcast.mapstore.GenericMapStore</class-name>
<properties>
<property name="data-connection-ref">my-mysql-database</property>
<property name="type-name">org.example.Person</property>
</properties>
</map-store>
</hazelcast>
single-column-as-valueIf set to true and the table has only one column (besides the identifier column), the value of this column will be used as the value in the IMap, instead of GenericRecord with only one field.false
xml
<hazelcast>
<map name="myMapName">
<map-store enabled="true">
<class-name>com.hazelcast.mapstore.GenericMapStore</class-name>
<properties>
<property name="data-connection-ref">my-mysql-database</property>
<property name="single-column-as-value">true</property>
</properties>
</map-store>
</hazelcast>
load-all-keysIf set to true, all keys will be read during MapLoader initialization.true
xml
<hazelcast>
<map name="myMapName">
<map-store enabled="true">
<class-name>com.hazelcast.mapstore.GenericMapStore</class-name>
<properties>
<property name="data-connection-ref">my-mysql-database</property>
<property name="load-all-keys">true</property>
</properties>
</map-store>
</hazelcast>

Supported backends​

The generic MapStore needs a SQL Connector that supports SELECT, UPDATE, SINK INTO and DELETE statements.

Officially supported connectors:

  • MySQL, PostgreSQL, Microsoft SQL Server, Oracle (it uses JDBC SQL Connector).
  • MongoDB (make sure you have hazelcast-jet-mongodb artifact included on the classpath).

Next Steps​

See the MapStore configuration guide for details about configuration options, including caching behaviors.