PyMondrian

PyMondrian is a Python schema an DW generator for the Mondrian OLAP engine.

With PyMondrian you create Python objetcs that are mapped into a Mondrian schema, that is, schema, cubes, dimensions, hierarchies, levels, annotations, meassures, etc.

PyMondrian also generates the DDL SQL scripts to create the datawarehouse tables.

The XML schema can be generated to a file or returned as a string.

The generated schema is based on Mondrian 3 according to the XML model for Mondrian schemas available at http://mondrian.pentaho.com/documentation/xml_schema.php.

Usage

A working example is available at https://github.com/gabitoju/pymondrian/blob/master/example.py.

This examples creates and generates a simple schema with 4 dimensions (Date, Customer, Supplier, Product) and 5 measures (Quantity, Subtotal, Taxes, Total, Subtotal Average).

The data.sql file contains sample data to load into the tables so you can test the schema with an OLAP client such as Saiku o JPivot.

The Mondrian schema and the DDL can be generated to a file or to the stdout.

from pymondrian.schema import *
from pymondrian.generator import generator
from pymondrian.mysql_generator import MySQLGenerator
    
# Create the schema, cubes, dimensions, hierarchies, levels...
    
# To stdout 
print generate(schema, output=1)

# To a file with the schema name as file name
generate(schema)

# To a file with other name
generate(schema, file_name='file.xml')

# Generate the DDL to stdout
mysql_generator = MySQLGenerator()
print mysql_generator.generate(schema, output=1)

# To a file with the schema name as file name
mysql_generator.generate(schema)

# To a file with other name
mysql_generator.generate(schema, file_name='file.sql')

Current version features

PyMondrian 0.2 generates the most basic schema in order to have a working OLAP Mondrian cube.

The available elements are:

• Schema
• Annotation
• Table
• Dimension
• Level
• Meassure

Basic DDL SQL generation is supported for MySQL via the MySQLGenerator. It generates the basic tables (dimension tables and fact table) but it has some limitations:

• It uses only one hierarchy.
• If a schema has two cubes and thos cubes share a dimensions, the DDL for the dimensions is generated twice.

Using not implemented attributes

Even though not all the attributes of the available elements have been implemented, it is possible to add them in a very simple way.

This code snippet shows how to do that:

# Import the Attribute class
from pymondrian.core.attribute import Attribute

# Let's add the formatString to a meassure
format_string = Attribute(xml_attribute_name='formatString', value='####')

# Now we assign it to the meassure like this
cuantity._format_string = format_string

You can name the attribute whatever way you like it but the following rules must be respected:

• The xml_attribute_name property should be named as the Mondrian attribute is named
• The property that contains the Attribute instance must start with '_'

API

Todo

• Implement Calculated Member and Key Expressions
• Generate API documentation
• Improve documentation
• Implement remaining elements
• Implement DDL SQL generator
• Implement all the attributes of the current available elements