Red Hat Training

A Red Hat training course is available for Red Hat Fuse

Chapter 79. JDBC

JDBC Component

The JDBC component enables you to access databases through JDBC, where SQL queries (SELECT) and operations (INSERT, UPDATE, and so on) are sent in the message body. This component uses the standard JDBC API, unlike the SQL Component component, which uses spring-jdbc.
Warning
This component can only be used to define producer endpoints, which means that you cannot use the JDBC component in a from() statement.

URI format

jdbc:dataSourceName[?options]
This component only supports producer endpoints.
You can append query options to the URI in the following format, ?option=value&option=value&...

Options

Name Default Value Description
readSize 0 The default maximum number of rows that can be read by a polling query.
statement.<xxx> null Apache Camel 2.1: Sets additional options on the java.sql.Statement that is used behind the scenes to execute the queries. For instance, statement.maxRows=10. For detailed documentation, see the java.sql.Statement javadoc documentation.
useJDBC4ColumnNameAndLabelSemantics true Sets whether to use JDBC 4/3 column label/name semantics. You can use this option to turn it false in case you have issues with your JDBC driver to select data. This only applies when using SQL SELECT using aliases (e.g. SQL SELECT id as identifier, name as given_name from persons).
resetAutoCommit true Camel 2.9: If true, Camel will set the autoCommit on the JDBC connection to be false, commit the change after executing the statement and reset the autoCommit flag of the connection at the end. If the JDBC connection does not support resetting the autoCommit flag, set this to false. When used with XA transactions you most likely need to set it to false so that the transaction manager is in charge of committing this tx.
allowNamedParameters true Camel 2.12: Whether to allow using named parameters in the queries.
prepareStatementStrategy Camel 2.12: Allows to plugin to use a custom org.apache.camel.component.jdbc.JdbcPrepareStatementStrategy to control preparation of the query and prepared statement.
useHeadersAsParameters false Camel 2.12: Set this option to true to use the prepareStatementStrategy with named parameters. This allows to define queries with named placeholders, and use headers with the dynamic values for the query placeholders.
outputType SelectList
Camel 2.12.1: Make the output of the producer to SelectList as List of Map, or SelectOne as single Java object in the following way: a) If the query has only single column, then that JDBC Column object is returned. (such as SELECT COUNT( * ) FROM PROJECT will return a Long object. b) If the query has more than one column, then it will return a Map of that result. c) If the outputClass is set, then it will convert the query result into an Java bean object by calling all the setters that match the column names. It will assume your class has a default constructor to create instance with. From Camel 2.14 onwards, SelectList is also supported. d) If the query resulted in more than one rows, it throws an non-unique result exception.
Camel 2.14.0: New StreamList output type value that streams the result of the query using an Iterator<Map<String, Object>>, it can be used along with the Splitter EIP.
outputClass null Camel 2.12.1: Specify the full package and class name to use as conversion when outputType=SelectOne. From Camel 2.14 onwards then SelectList is also supported.
beanRowMapper Camel 2.12.1: To use a custom org.apache.camel.component.jdbc.BeanRowMapper when using outputClass. The default implementation will lower case the row names and skip underscores, and dashes. For example "CUST_ID" is mapped as "custId".
useGetBytesForBlob false Camel 2.16: To read BLOB columns as bytes instead of string data. This may be needed for certain databases such as Oracle where you must read BLOB columns as bytes.

Result

By default, the result is returned in the OUT body as an ArrayList<HashMap<String, Object>>. The List object contains the list of rows and the Map objects contain each row with the String key as the column name.
Note
This component fetches ResultSetMetaData to be able to return the column name as the key in the Map.

Message Headers

Header Description
CamelJdbcRowCount If the query is a SELECT, the row count is returned in this OUT header.
CamelJdbcUpdateCount If the query is an UPDATE, the update count is returned in this OUT header.
CamelGeneratedKeysRows Camel 2.10: Rows that contains the generated kets.
CamelGeneratedKeysRowCount Camel 2.10: The number of rows in the header that contains generated keys.
CamelJdbcColumnNames Camel 2.11.1: The column names from the ResultSet as a java.util.Set type.
CamelJdbcParametes Camel 2.12: A java.util.Map which has the headers to be used if useHeadersAsParameters has been enabled.

Generated keys

Available as of Camel 2.10
If you insert data using SQL INSERT, then the RDBMS may support auto generated keys. You can instruct the JDBC producer to return the generated keys in headers. To do that set the header CamelRetrieveGeneratedKeys=true. Then the generated keys will be provided as headers with the keys listed in the table above.
from("direct:insert")
  .setHeader("CamelRetrieveGeneratedKeys", constant(true))
  .setBody(constant("INSERT INTO projects (project) VALUES ('Camel')"))
  .to("jdbc:myDataSource");
You can see more details in this unit test.
Important
Using generated keys does not work with together with named parameters.

Using named parameters

Available as of Camel 2.12
In the given route below, we want to get all the projects from the projects table. Notice the SQL query has 2 named parameters, :?lic and :?min. Camel will then lookup these parameters from the message headers. Notice in the example above we set two headers with constant value for the named parameters:
  from("direct:projects")
     .setHeader("lic", constant("ASF"))
     .setHeader("min", constant(123))
     .setBody(constant("select * from projects where license = :?lic and id > :?min order by id"))
     .to("jdbc:myDataSource?useHeadersAsParameters=true")
You can also store the header values in a java.util.Map and store the map on the headers with the key CamelJdbcParameters.

Samples

In the following example, we fetch the rows from the customer table.
First we register our datasource in the Apache Camel registry as testdb:
JndiRegistry reg = super.createRegistry();
reg.bind("testdb", db);
return reg;
Then we configure a route that routes to the JDBC component, so the SQL will be executed. Note how we refer to the testdb datasource that was bound in the previous step:
// lets add simple route
public void configure() throws Exception {
    from("direct:hello").to("jdbc:testdb?readSize=100");
}
Or you can create a DataSource in Spring like this:
<camelContext id="camel" xmlns="http://camel.apache.org/schema/spring">
  <route>
	 <!-- trigger every second -->
     <from uri="timer://kickoff?period=1s"/>
     <setBody>
       <constant>select * from customer</constant>
     </setBody>
     <to uri="jdbc:testdb"/>
     <to uri="mock:result"/>
  </route>
</camelContext>
<!-- Just add a demo to show how to bind a date source for camel in Spring-->
<jdbc:embedded-database id="testdb" type="DERBY">
	<jdbc:script location="classpath:sql/init.sql"/>
</jdbc:embedded-database>
We create an endpoint, add the SQL query to the body of the IN message, and then send the exchange. The result of the query is returned in the OUT body:
// first we create our exchange using the endpoint
Endpoint endpoint = context.getEndpoint("direct:hello");
Exchange exchange = endpoint.createExchange();
// then we set the SQL on the in body
exchange.getIn().setBody("select * from customer order by ID");

// now we send the exchange to the endpoint, and receives the response from Camel
Exchange out = template.send(endpoint, exchange);

// assertions of the response
assertNotNull(out);
assertNotNull(out.getOut());
List<Map<String, Object>> data = out.getOut().getBody(List.class);
assertNotNull(data);
assertEquals(3, data.size());
Map<String, Object> row = data.get(0);
assertEquals("cust1", row.get("ID"));
assertEquals("jstrachan", row.get("NAME"));
row = data.get(1);
assertEquals("cust2", row.get("ID"));
assertEquals("nsandhu", row.get("NAME"));
If you want to work on the rows one by one instead of the entire ResultSet at once, you need to use the Splitter EIP such as:
from("direct:hello")
// here we split the data from the testdb into new messages one by one
// so the mock endpoint will receive a message per row in the table
// the StreamList option allows to stream the result of the query without creating a List of rows
// and notice we also enable streaming mode on the splitter
.to("jdbc:testdb?outputType=StreamList")
  .split(body()).streaming()
  .to("mock:result");

Sample - Polling the database every minute

If we want to poll a database using the JDBC component, we need to combine it with a polling scheduler such as the Timer or Quartz etc. In the following example, we retrieve data from the database every 60 seconds:
from("timer://foo?period=60000").setBody(constant("select * from customer")).to("jdbc:testdb").to("activemq:queue:customers");

Sample - Move Data Between Data Sources

A common use case is to query for data, process it and move it to another data source (ETL operations). In the following example, we retrieve new customer records from the source table every hour, filter/transform them and move them to a destination table:
from("timer://MoveNewCustomersEveryHour?period=3600000")
    .setBody(constant("select * from customer where create_time > (sysdate-1/24)"))
    .to("jdbc:testdb")
    .split(body())
		.process(new MyCustomerProcessor()) //filter/transform results as needed
        .setBody(simple("insert into processed_customer values('${body[ID]}','${body[NAME]}')"))
        .to("jdbc:testdb");

See also