sjtufighter/optiq

Dynamic data management framework

optiq

Optiq is a dynamic data management framework.

Prerequisites

Optiq requires git, maven (3.0.4 or later), and JDK 1.6 or later (JDK 1.7 preferred).

Download and build

$ git clone git://github.com/julianhyde/optiq.git
$ cd optiq
$ mvn install

Use mvn -DskipTests if you do not want to execute the tests.

Example

Optiq makes data anywhere, of any format, look like a database. For example, you can execute a complex ANSI-standard SQL statement on in-memory collections:

public static class HrSchema {
  public final Employee[] emps = ... ;
  public final Department[] depts = ...;
}

Class.forName("net.hydromatic.optiq.jdbc.Driver");
Connection connection = DriverManager.getConnection("jdbc:optiq:");
OptiqConnection optiqConnection =
    connection.unwrap(OptiqConnection.class);
ReflectiveSchema.create(optiqConnection,
    optiqConnection.getRootSchema(), "hr", new HrSchema());
Statement statement = optiqConnection.createStatement();
ResultSet resultSet = statement.executeQuery(
    "select d.\"deptno\", min(e.\"empid\")\n"
    + "from \"hr\".\"emps\" as e\n"
    + "join \"hr\".\"depts\" as d\n"
    + "  on e.\"deptno\" = d.\"deptno\"\n"
    + "group by d.\"deptno\"\n"
    + "having count(*) > 1");
print(resultSet);
resultSet.close();
statement.close();
connection.close();

Where is the database? There is no database. The connection is completely empty until ReflectiveSchema.create registers a Java object as a schema and its collection fields emps and depts as tables.

Optiq does not want to own data; it does not even have favorite data format. This example used in-memory data sets, and processed them using operators such as groupBy and join from the linq4j library. But Optiq can also process data in other data formats, such as JDBC. In the first example, replace

ReflectiveSchema.create(optiqConnection,
    optiqConnection.getRootSchema(), "hr", new HrSchema());

with

Class.forName("com.mysql.jdbc.Driver");
BasicDataSource dataSource = new BasicDataSource();
dataSource.setUrl("jdbc:mysql://localhost");
dataSource.setUsername("sa");
dataSource.setPassword("");
JdbcSchema.create(optiqConnection, dataSource, rootSchema, "hr", "");

and Optiq will execute the same query in JDBC. To the application, the data and API are the same, but behind the scenes the implementation is very different. Optiq uses optimizer rules to push the JOIN and GROUP BY operations to the source database.

In-memory and JDBC are just two familiar examples. Optiq can handle any data source and data format. To add a data source, you need to write an adapter that tells Optiq what collections in the data source it should consider "tables".

For more advanced integration, you can write optimizer rules. Optimizer rules allow Optiq to access data of a new format, allow you to register new operators (such as a better join algorithm), and allow Optiq to optimize how queries are translated to operators. Optiq will combine your rules and operators with built-in rules and operators, apply cost-based optimization, and generate an efficient plan.

Non-JDBC access

Optiq also allows front-ends other than SQL/JDBC. For example, you can execute queries in linq4j:

final OptiqConnection connection = ...;
ParameterExpression c = Expressions.parameter(Customer.class, "c");
for (Customer customer
    : connection.getRootSchema()
        .getSubSchema("foodmart")
        .getTable("customer", Customer.class)
        .where(
            Expressions.<Predicate1<Customer>>lambda(
                Expressions.lessThan(
                    Expressions.field(c, "customer_id"),
                    Expressions.constant(5)),
                c))) {
  System.out.println(c.name);
}

Linq4j understands the full query parse tree, and the Linq4j query provider for Optiq invokes Optiq as an query optimizer. If the customer table comes from a JDBC database (based on this code fragment, we really can't tell) then the optimal plan will be to send the query

SELECT *
FROM "customer"
WHERE "customer_id" < 5

to the JDBC data source.

Writing an adapter

The optiq-csv project provides a CSV adapter, which is fully functional for use in applications but is also simple enough to serve as a good template if you are writing your own adapter.

See the optiq-csv tutorial for information on using optiq-csv and writing adapters.

See the HOWTO for more information about using other adapters, and about using Optiq in general.

Status

The following features are complete.

• Query parser, validator and optimizer
• Support for reading models in JSON format
• Many standard functions and aggregate functions
• JDBC queries against Linq4j and JDBC back-ends
• Linq4j front-end
• SQL features: SELECT, FROM (including JOIN syntax), WHERE, GROUP BY (and aggregate functions including COUNT(DISTINCT ...)), HAVING, ORDER BY (including NULLS FIRST/LAST), set operations (UNION, INTERSECT, MINUS), sub-queries (including correlated sub-queries), windowed aggregates, LIMIT (syntax as Postgres)

For more details, see the Reference guide.

Drivers

• JDBC driver

Adapters

• Apache Drill adapter
• Cascading adapter (Lingual)
• CSV adapter (optiq-csv)
• JDBC adapter (part of optiq-core)
• MongoDB adapter (optiq-mongodb)
• Spark adapter (optiq-spark)
• Splunk adapter (optiq-splunk)
• Eclipse Memory Analyzer (MAT) adapter (optiq-mat-plugin)

More information

• License: Apache License, Version 2.0.
• Author: Julian Hyde
• Blog: http://julianhyde.blogspot.com
• Incubation status page: http://incubator.apache.org/projects/optiq.html
• Project page: http://www.hydromatic.net/optiq
• Source code: http://github.com/julianhyde/optiq
• Issues: Apache JIRA
• Developers list: dev at optiq.incubator.apache.org (archive, subscribe)
• HOWTO
• JSON model
• Reference guide
• Release notes and history

Pre-Apache resources

These resources, which we used before Optiq was an Apache incubator project, are for reference only. They may be out of date. Please don't post to the mailing list.

• Developers list: [email protected]

Presentations

• How to integrate Splunk with any data solution (Splunk User Conference, 2012)
• Drill / SQL / Optiq (2013)
• SQL on Big Data using Optiq (2013)
• SQL Now! (NoSQL Now! conference, 2013)
• Cost-based optimization in Hive (Hadoop Summit, 2014)
• Discardable, in-memory materialized query for Hadoop (Hadoop Summit, 2014)