Microsoft Docs Parser

When I decided to post my T-SQL script library on this site,  I wanted to be able to identify things like T-SQL functions, data types and language elements along with SQL Server system objects (i.e. system views, system procedures, system tables, system functions) that were in each script.

The Microsoft Docs Parser is a program I developed to parse information from Microsoft’s publicly available Transact-SQL and SQL Server documentation. The information populated a SQL Server database used to generate the reference information you’ll find under every script in the T-SQL section of this website. It’s also used to create the list of T-SQL and SQL Server keywords and phrases that were used to add search functionality to the site and search engine optimization to each page.

Enabling Search Functionality

One reason for posting my T-SQL scripts on this website was to make the library searchable.  The library consists of queries and scripts that were stored in my snippet manager software’s XML data file and in .sql files on my hard drive.  Combining them into a SQL Server database and then posting them to my website provided the ability to search all of my scripts, regardless of where they’re stored.

None of this is helpful if no one can find my website, so basic search engine optimization is required. That meant having the ability to search for similar scripts.  It also meant providing search engines with the information they look for so that the scripts can be indexed appropriately.

Gathering Reference Information

This project created a database containing the following reference info about each keyword or object:

1. Description of the object
2. URL to documentation
3. URL to GitHub repository
4. Related keywords
5. See also information
6. Server version the documentation belongs to
7. Columns returned (table valued functionst/table)

Search Engine Optomization

For this site to be useful, people need to find it, so I needed to think about what the way I search for the information that’s on this site and then optimize each page so that search engines see it that way.

I typically search for code snippets a couple of different ways. One way is to search for what I want the script to do.  If they’re like me, readers will arrive at my site after searching for a functional goal, such as “best way to store C# enum in T-SQL.”

If I already know what object I want to use but need to see an example of its usage, I typically search for an object using its name and language, like this:  “tsql CHOOSE example”.

To position my scripts well for both types of searches, I made sure that the title of each script represented the functional goal it achieves.  Then, I used another application I developed called T-SQL to WordPress HTML  to parse out the keywords and object names from my T-SQL scripts for use as keywords in the web page’s metadata.