Home > Articles > Data > SQL Server

SQL Server Reference Guide

Hosted by

Microsot SQL Server Reporting Services

Last updated Mar 28, 2003.

When you think about it, there are really just three functions in enterprise computing: processing, storage and reporting. That may sound a bit oversimplified, but in fact your users have various methods for creating and manipulating data (processing), the data they create is retained in many formats (storage), and the users retrieve the data in yet other formats for presentation (reporting). Of these three, the one that is least familiar to many programmers and DBAs is reporting. But reporting is where most users see the results of the other two functions, and often the source of the greatest value for the business. You can combine all of your enterprise data in a single report, showing data from multiple functions from the organization. With a coherent reporting infrastructure, you can store data authoritatively one time, and combine it to provide a single view. But there are challenges when you try and bring this data together, especially if you use a production On Line Transaction Processing system (OLTP) as the reporting base data.

Data-entry applications (OLTP) are made to allow fast inserts, updates and deletes for production databases. These actions use locks, to protect data integrity. However, when users want to report on that data, locks can slow down the reports, or even prevent them from being generated. Inserts, updates and deletes can get in the way while reports are accessing data, affecting report integrity. This is rarely a problem if the user base or activity is small, or if the data used for reports isn’t the same as the base data for those used for reporting.

In large enterprises, however, this is a critical issue. Automated line entry machines or high volumes of data entry demand rapid access to the database. Reporting-caused slowdowns on these changes affect the bottom line. At the same time, line quality systems, managers, and other audiences require complex, timely reports that access a lot of data, often for a long time.

To balance these two needs, most large applications use a special reporting infrastructure, separate from the main application. The application uses separate menus and screens for the users, as well as separate servers and — in some cases — separate databases.

If your shop develops the application, you might use a third-party system, such as Crystal Enterprise. Crystal Enterprise (and its competitors) use a series of services, servers, and databases to provide data to several output mechanisms, including a report design and display tool called Crystal Reports.

Microsoft has a similar set of tools, called Reporting Services. Originally designed for SQL Server 2005, Microsoft released it to fill a user need and to stay competitive with Oracle and IBM (Oracle and IBM have report solutions which you can purchase for their platforms). Reporting Services comes built-into the SQL Server Editions above Express, so you don’t need a separate license for it unless you put it on another server (physical or virtual).

Using Reporting Services, you can create a full-fledged reporting infrastructure, but it does take planning. In this overview I’ll explain the various parts of this feature and how you can put it to use in your organization. Even if you don’t develop software for a living, you can use Reporting Services.

Reporting Services components are divided into various parts: the Report Server, with its Report Server Database; an HTTP server, a report creation tool (there are two of these) and the Report Manager, a .NET-based web interface for management of the reports. All of these parts access data of some sort — and here’s where the real power of this feature comes into play — Reporting Services can hit data sources from Oracle, SQL Server, Excel, Text files, and many more. You can query a piece of data from one system and put it on the same report as a datum from another source. Your users will see the output in a single display.

From the “Start” menu on your Reporting Services server, you’ll find a new tool called the “Reporting Services Configuration Manager”. This web-based tool is how the administrator sets up the sites that SQL Server will use for Reports, and also security access for those sites. I’ll cover the complete configuration setup in another tutorial.

There are multiple ways for you or your users to create a report. The first is to use the Business Information Development Studio, or BIDS, that is installed when you install the SQL Server client tools. This is a full development environment that allows you to work “offline” from the Reporting Services system, and once you’re done you can “publish” the report to a Reporting Services instance.

You can also create a new report using a web-based tool called the Report Builder that is much lighter and easier to work with that BIDS. You can also allow your users to work with this tool, even limiting what data they can work with. This tool has been updated, so make sure you visit the links at the end of this overview to learn where to get the latest one, and which version of SQL Server Reporting Services it works with.

You can also use various third-party tools. In fact, all of these tools end up creating an XML-like text file called Report Definition Language (RDL). Many tools support this definition for report design, so you can create reports in one tool and export it to another. The RDL is actually what gets sent to the Reporting Services server, which in turn processes the file, gets the data, and displays it using an HTTP or HTTPS transport.

With the report designed, you publish the report definition to the web component of Reporting Services. You manage the report distribution from Reporting Services Configuration Manager, which also handles security.

The report database contains the meta-data for the report, and you can use data from lots of sources, such as Oracle databases, SQL Server databases, and Excel spreadsheets. Any or all of these datasets (not to be confused with .NET datasets) can exist on a single report. The reports support many outputs, including PDF, Web Archive, TIFF, HTML, Excel, CSV, and XML.

You can also schedule the reports to run at a later time. You can also access the reports via a "pull" mechanism, using an API. Reports can also be accessed through FTP or sent by e-mail. Reports can be parameterized and also be event-driven.

Once a report is created, its definition is stored in the Reporting Services database. This is a separate database from where the data the report access is stored – it’s the “meta data” that the Reporting Services system needs to run.

The reporting services engine, called the Report Processor, receives a request for a report from a web page hosted either on the Reporting Server, or from another web server. This request includes parameters and formatting. In SQL Server 2005, Reporting Services requires Internet Information Services (IIS) to be installed on the Reporting Services Instance, but starting with SQL Server 2008 this requirement was removed, and SQL Server now hosts its own web service interface.

The Report Processor gets the report definition from the request using the Report Definition Language, and obtains the report data from the data sources. Inside the RDL of the report you’ve put in the information to get that data, and there are a few ways to handle the security for the remote data source. You can embed the security, so that the report accesses the data on behalf of the user, or you can use the user’s current security context to pass along to the data source. There are other methods you can use for security, and it’s important to learn the ramifications for each one, as each situation will be different.

The Report Processor then transforms the data and sends it along with a schema to the rendering engine, called the Rendering Extension.

Finally, your users can access the reports right over the web, simply by pointing to the Reporting Services server address for either a single report or a group of them. You can also publish a report out to a file, even in other formats, and have the users pick them up from there or you can schedule another process to e-mail them or post them to an FTP location. Another option is SharePoint — if you have the latest versions of that product, you can “publish” reports from the Reporting Services server to the SharePoint sites of your choice. As you can see, you have a lot of choices for the distribution of the reports.

Now that you understand the architecture of a Reporting Services system, I’ll show you how to create a simple report that you can use for yourself, so that you’re able to make others for your users.

One final item: you get a license for Reporting Services in each installation of SQL Server (except for the free versions, such as Express). If you choose to split out the report server from the production database (which I highly advise), you’ll need another SQL Server license to cover that.

InformIT Articles and Sample Chapters

I cover Report Builder in more detail in this Reference Guide entry, The SQL Server Central Management System: Reporting the Data and Project Summary, which is part of a larger series on monitoring your servers. Learning how to make reports for yourself is a great way to learn how to make them for others, and this tool is quite simple to use.

Books and eBooks

Michael Lisin, Jim Joseph and Amit Goyal cover this topic very well in their book Microsoft SQL Server 2008 Reporting Services Unleashed, which you can also get as an eBook or in Safari Books Online.

Online Resources

The Report Builder tool is linked here — make sure you check the links at the bottom of that page to get the latest version.