Skip to content

Latest commit

 

History

History
116 lines (81 loc) · 2.96 KB

ideas.md

File metadata and controls

116 lines (81 loc) · 2.96 KB
Raw

OLAP.js Ideas

Installation:

	> npm install olap

Run stand-alone server:

	> olap

Create stand-alone server in Node.js:

	var olap = require('olap');
	olap.server({port:3000, xmla:'/xmla'});

Embed server into Express framework with standard path ('/xmla'):

	var express = require('express');
	var app = express();
	var olap = require('olap');

	app.use(olap.express());

Embed server into Express framework with non-standard paths:

	app.get('/discover',function(req,res){
		res.send(olap.discover());
	});

or

	app.use(olap.express({xmla:'/xmla1'}));

Server entry points:

XML/A:

	olap.xmla(xmla);
	// returns XML formed answer

Simple JSON interface:

	olap.discover(parms);
	// returns array of cubes

	olap.execute(mdx_statement,params);
	// returns recordset object

	olap.ToXML(recordset);
	// returns XML object for XML/A answer

Also there are some functions for creation and modification of:

  • datasources
  • schemas
  • cubes
  • dimensions
  • measures

How the server works

  1. Client (e.g. MS Excel or HTML pivot control) sends XML/A request. Usually, the client sends some DISCOVERY queries, and then one EXECUTE query. This is a POST message with XML/A text in the body (for Execute query it includes MDX statement inside).

  2. Server catch the path, unpack XML/A query and run olap.xmla()

  3. This function check the type of query, and if it is DISCOVERY then it runs olap.discovery() method, and olap.execute() in opposite case.

  4. olap.discovery() returns information about cudes, measures, etc.

  5. olap.execute() parses MDX statement to AST (abstract syntax tree). then it calls olap.query(mdx_ast) function.

  6. olap.query() translate MDX to the set of SQL statements and run them (this is the most intelligent part of the server). Some SQL queries can be cached in the memory,

  7. After SQL queries Olap.js will run post processing (this operation can be paralleled with intelligent map-reduce algorithm). Here Olap.js will make all aggregations.

  8. The result set is packed into JSON and then can be packed into XML envelope and return to the client.

MDX Parser

We will use the similar parser from AlaSQL and AlaMDX.

Data Sources

We will use stadard connectors to Postgres, MySQL, SQLite, and MSSQL Server. We will also use AlaSQL for fast tests and simple debugging.

OLAP Processor

The most intelligent part of the system.

The first queries request for Dimensions. After that Olap.js will construct query for Measures and groupping data. After that Olap.js will postprocess data (construct rows and columns queries).

Cube Editor

OLAP.js will save data about cubes, dimensions and measures in JSON format and saves it to the disk. Later we will create visual editor (like SSAS and Mondrian do).