SQL

Most NoSQL products employ a custom query language. In this, OrientDB differs by focusing on standards in query languages. That is, instead of inventing “Yet Another Query Language,” it begins with the widely used and well-understood language of SQL. It then extends SQL to support more complex graphing concepts, such as Trees and Graphs.

Why SQL? Because SQL is ubiquitous in the database development world. It is familiar and more readable and concise than its competitors, such as Map Reduce scripts or JSON based querying.

SELECT

The SELECT statement queries the database and returns results that match the given parameters. For instance, earlier in Getting Started, two queries were presented that gave the same results: BROWSE CLUSTER ouser and BROWSE CLASS OUser. Here is a third option, available through a SELECT statement.

  1. orientdb> SELECT FROM OUser

Notice that the query has no projections. This means that you do not need to enter a character to indicate that the query should return the entire record, such as the asterisk in the Relational model, (that is, SELECT * FROM OUser).

Additionally, OUser is a class. By default, OrientDB executes queries against classes. Targets can also be:

  • Clusters To execute against a cluster, rather than a class, prefix CLUSTER to the target name.

    1. orientdb> SELECT FROM CLUSTER:Ouser
  • Record ID To execute against one or more Record ID’s, use the identifier(s) as your target. For example.

    1. orientdb> SELECT FROM #10:3
    2. orientdb> SELECT FROM [#10:1, #10:30, #10:5]
  • Indexes To execute a query against an index, prefix INDEX to the target name.

    1. orientdb> SELECT VALUE FROM INDEX:dictionary WHERE key='Jay'

WHERE

Much like the standard implementation of SQL, OrientDB supports WHERE conditions to filter the returning records too. For example,

  1. orientdb> SELECT FROM OUser WHERE name LIKE 'l%'

This returns all OUser records where the name begins with l. For more information on supported operators and functions, see WHERE.

ORDER BY

In addition to WHERE, OrientDB also supports ORDER BY clauses. This allows you to order the results returned by the query according to one or more fields, in either ascending or descending order.

  1. orientdb> SELECT FROM Employee WHERE city='Rome' ORDER BY surname ASC, name ASC

The example queries the Employee class, it returns a listing of all employees in that class who live in Rome and it orders the results by surname and name, in ascending order.

GROUP BY

In the event that you need results of the query grouped together according to the values of certain fields, you can manage this using the GROUP BY clause.

  1. orientdb> SELECT SUM(salary) FROM Employee WHERE age < 40 GROUP BY job

In the example, you query the Employee class for the sum of the salaries of all employees under the age of forty, grouped by their job types.

LIMIT

In the event that your query returns too many results, making it difficult to read or manage, you can use the LIMIT clause to reduce it to the top most of the return values.

  1. orientdb> SELECT FROM Employee WHERE gender='male' LIMIT 20

In the example, you query the Employee class for a list of male employees. Given that there are likely to be a number of these, you limit the return to the first twenty entries.

SKIP

When using the LIMIT clause with queries, you can only view the topmost of the return results. In the event that you would like to view certain results further down the list, for instance the values from twenty to forty, you can paginate your results using the SKIP keyword in the LIMIT clause.

  1. orientdb> SELECT FROM Employee WHERE gender='male' LIMIT 20
  2. orientdb> SELECT FROM Employee WHERE gender='male' SKIP 20 LIMIT 20
  3. orientdb> SELECT FROM Employee WHERE gender='male' SKIP 40 LIMIT 20

The first query returns the first twenty results, the second returns the next twenty results, the third up to sixty. You can use these queries to manage pages at the application layer.

INSERT

The INSERT statement adds new data to a class and cluster. OrientDB supports three forms of syntax used to insert new data into your database.

  • The standard ANSI-93 syntax:

    1. orientdb> INSERT INTO Employee(name, surname, gender)
    2. VALUES('Jay', 'Miner', 'M')
  • The simplified ANSI-92 syntax:

    1. orientdb> INSERT INTO Employee SET name='Jay', surname='Miner', gender='M'
  • The JSON syntax:

    1. orientdb> INSERT INTO Employee CONTENT {name : 'Jay', surname : 'Miner',
    2. gender : 'M'}

Each of these queries adds Jay Miner to the Employee class. You can choose whichever syntax that works best with your application.

UPDATE

The UPDATE statement changes the values of existing data in a class and cluster. In OrientDB there are two forms of syntax used to update data on your database.

  • The standard ANSI-92 syntax:

    1. orientdb> UPDATE Employee SET local=TRUE WHERE city='London'
  • The JSON syntax, used with the MERGE keyword, which merges the changes with the current record:

    1. orientdb> UPDATE Employee MERGE { local : TRUE } WHERE city='London'

Each of these statements updates the Employee class, changing the local property to TRUE when the employee is based in London.

DELETE

The DELETE statement removes existing values from your class and cluster. OrientDB supports the standard ANSI-92 compliant syntax for these statements:

  1. orientdb> DELETE FROM Employee WHERE city <> 'London'

Here, entries are removed from the Employee class where the employee in question is not based in London.

See also: