Saturday, February 9, 2013

ODATA for Dummies

In a world full of open source and open standards what's so important about one more protocol or specification, right? Why should you know about it, right? Right, but technology professionals are required to know a lot more technologies today than they were in the past because of the rapidly shifting software landscape. Gone are the days when you could master one technology and keep your job forever. At any moment you could be asked to look into some new technology and be expected to work on it or create a proof of concept in an effort to evaluate its fit to your current project. ODATA may be one such technology you may not have worked in or heard of, but worth knowing about.

If you already know about ODATA specification and how it works, please stop reading. This is a very high level but useful information on ODATA meant for newcomers.

Promoted by Microsoft ( do I hear Microsoft haters leaving already? ) ODATA is a protocol to access information over the web. ODATA defines a standard way of exposing any information through  industry standard HTTP protocol. It also follows the architecture popularly known as REST. You can get details of the protocol and much more here. If you want to expose your data to your clients using ODATA, you will need much deeper understanding of the specification. However if you just want to consume the data as a client, you could be up and running in a few minutes to a few hours depending on your level of expertise. This post will have you create a sample and test it within minutes.

Let's take an example of a company that we all know about - Netflix. It uses odata service to publish its movie catalog. Viewing the movie catalog or searching for a title could easily be done by just accessing certain urls through your browser. You just need to know how to build further urls based on the information you get from a url. ODATA provides the response to your http request in a standard XML format that follows either Atom or Json format. Try these urls yourself and see how easy it is get to the information. The odata services are consumed through  urls that are intuitive and easy to understand. will show what type of information you could get from this website. The URL will resolve to The results will show that you could get a list of Genres, Titles and Languages. Now, suppose you want to see all the titles Netflix has, then simply add 'Titles' to the previous url as- This will generate all the titles and associated information. If you want to know all the details (metadata) about all the entities (collections) simply type in this url.$metadata.This results in information on collections, the entities they are made up of and the attributes of each entity there in. This just shows you how easy it is to get to this basic information. Now you are ready to dig deeper.

The data that shows up on your screen may have been stored in a table in a relational database or  in an xml database or in a a file in the file system, it does not matter. The information is available to you through the http get method (browser sends 'get' method to the end point). Irrespective of the type and location of the data storage, specific data is likely identified by a unique key. If you know the unique identifier of the entity you want to access, then you could use it to get to more information.  The following example shows how e Titles are uniquely identified by an 'id' field. This information - that id is the unique identifier of a title is gotten from the previous metadata query we executed. Spend a few minutes reviewing the results of metadata query.'13bLK').

When you don't know the unique identifier you could always search  titles based on a name (name is an attribute of Title entity as revealed by the metadata query.)$filter=(Name%20eq%20'Rod%20Stewart')

The count of total number of titles Netflix owns is easily found by using the following url. As of 2/8/2013 the count is 160050. Odata defines many such functions count, length, replace, trim etc.$count

So this is what odata specification provides. It specifies the format of the url, the functions and operators you could use in the url to get to the data you want. REST based architecture exposes all the information through http metods - in this case - 'GET' is what we have seen so far.

Now a little bit on to how providers like Netflix create these services. If service provider is exposing mechanisms for creating, updating and deleting the data through odata producer and if the information that's being exposed is in relational database, they need to map the standard CRUD methods to http methods. So create is mapped to post, delete to delete, update to put and select to get. The essential tasks of the producer would be to parse the request, get the http method, get the content, call the back end where the data is stored, perform the mapped functions and convert the retrieved data to Atom or Json format before it's sent back to the client. Vendors like Microsoft and Google (odata4j) are offering function libraries that makes it easy to create your own producer. If you just want to consume the odata services there are many client libraries like popular datajs.

Now you are ready to create your own client.
Here is a snippet of code I got from datajs tutorial. Save this as any html document and use it in any modern browser. You will see all the genres of all the content from the Netflix. Before running the sample don't forget to download datajs-1.1.0.min.js directory from here and save it to your c:\datajs directory.

I hope this post will give you enough information to try things on your own.  I also hope to have another post some time later on the same subject. Odata is actively used by companies such as Microsoft, SAP, Netflix, Facebook and Ebay. Only time will tell if this standard becomes widely used technology or not, but it already has gathered enough following to attract our attention. website. Good luck. Your comments are welcome. Please feel free to share if you like it.