Geode/Gemfire OQL Interpreter for Apache Zeppelin

Overview

Name Class Description
%geode.oql GeodeOqlInterpreter Provides OQL environment for Apache Geode

This interpreter supports the Geode Object Query Language (OQL).
With the OQL-based querying language:

zeppelin-view

  • You can query on any arbitrary object
  • You can navigate object collections
  • You can invoke methods and access the behavior of objects
  • Data mapping is supported
  • You are not required to declare types. Since you do not need type definitions, you can work across multiple languages
  • You are not constrained by a schema

This Video Tutorial illustrates some of the features provided by the Geode Interpreter.

Create Interpreter

By default Zeppelin creates one Geode/OQL instance. You can remove it or create more instances.

Multiple Geode instances can be created, each configured to the same or different backend Geode cluster. But over time a Notebook can have only one Geode interpreter instance bound. That means you cannot connect to different Geode clusters in the same Notebook. This is a known Zeppelin limitation.

To create new Geode instance open the Interpreter section and click the +Create button. Pick a Name of your choice and from the Interpreter drop-down select geode.
Then follow the configuration instructions and Save the new instance.

Note: The Name of the instance is used only to distinguish the instances while binding them to the Notebook. The Name is irrelevant inside the Notebook. In the Notebook you must use %geode.oql tag.

Bind to Notebook

In the Notebook click on the settings icon in the top right corner. The select/deselect the interpreters to be bound with the Notebook.

Configuration

You can modify the configuration of the Geode from the Interpreter section. The Geode interpreter expresses the following properties:

Property Name Description Default Value
geode.locator.host The Geode Locator Host localhost
geode.locator.port The Geode Locator Port 10334
geode.max.result Max number of OQL result to display to prevent the browser overload 1000

How to use

Tip 1: Use (CTRL + .) for OQL auto-completion.

Tip 2: Always start the paragraphs with the full %geode.oql prefix tag! The short notation: %geode would still be able run the OQL queries but the syntax highlighting and the auto-completions will be disabled.

Create / Destroy Regions

The OQL specification does not support Geode Regions mutation operations. To create/destroy regions one should use the GFSH shell tool instead. In the following it is assumed that the GFSH is colocated with Zeppelin server.

%sh
source /etc/geode/conf/geode-env.sh
gfsh << EOF

 connect --locator=ambari.localdomain[10334]

 destroy region --name=/regionEmployee
 destroy region --name=/regionCompany
 create region --name=regionEmployee --type=REPLICATE
 create region --name=regionCompany --type=REPLICATE

 exit;
EOF

Above snippet re-creates two regions: regionEmployee and regionCompany. Note that you have to explicitly specify the locator host and port. The values should match those you have used in the Geode Interpreter configuration. Comprehensive list of GFSH Commands by Functional Area.

Basic OQL

%geode.oql
SELECT count(*) FROM /regionEmployee

OQL IN and SET filters:

%geode.oql
SELECT * FROM /regionEmployee
WHERE companyId IN SET(2) OR lastName IN SET('Tzolov13', 'Tzolov73')

OQL JOIN operations

%geode.oql
SELECT e.employeeId, e.firstName, e.lastName, c.id as companyId, c.companyName, c.address
FROM /regionEmployee e, /regionCompany c
WHERE e.companyId = c.id

By default the QOL responses contain only the region entry values. To access the keys, query the EntrySet instead:

%geode.oql
SELECT e.key, e.value.companyId, e.value.email
FROM /regionEmployee.entrySet e

Following query will return the EntrySet value as a Blob:

%geode.oql
SELECT e.key, e.value FROM /regionEmployee.entrySet e

Note: You can have multiple queries in the same paragraph but only the result from the first is displayed. [1], [2].

GFSH Commands From The Shell

Use the Shell Interpreter (%sh) to run OQL commands form the command line:

%sh
source /etc/geode/conf/geode-env.sh
gfsh -e "connect" -e "list members"

Apply Zeppelin Dynamic Forms

You can leverage Zeppelin Dynamic Form inside your OQL queries. You can use both the text input and select form parameterization features

%geode.oql
SELECT * FROM /regionEmployee e WHERE e.employeeId > ${Id}

Auto-completion

The Geode Interpreter provides a basic auto-completion functionality. On (Ctrl+.) it list the most relevant suggestions in a pop-up window.

Geode REST API

To list the defined regions you can use the Geode REST API:

http://<geode server hostname>phd1.localdomain:8484/gemfire-api/v1/
{
  "regions" : [{
    "name" : "regionEmployee",
    "type" : "REPLICATE",
    "key-constraint" : null,
    "value-constraint" : null
  }, {
    "name" : "regionCompany",
    "type" : "REPLICATE",
    "key-constraint" : null,
    "value-constraint" : null
  }]
}

To enable Geode REST API with JSON support add the following properties to geode.server.properties.file and restart:

http-service-port=8484
start-dev-rest-api=true