Awesome job so far! You've created your first template and imported the meta data. Now we can use that meta data to develop a subset template. 

  • Let's switch from Project settings to Development

The development section consists of the start table section and four lists with tables. To create a good subset template, we have to classify our imported data model and choose a good start table and start filter. You will learn all about classifications and start filters in the next chapters, but first let's give a short introduction how Subset works. 

DATPROF Subset uses a patented algorithm that use the foreign keys (relations) between tables to generate a dependency network to efficiently extract data from a single starting table. That means that foreign keys are required to extract a good subset. Sometimes your database will not have foreign keys stored within the database itself. In that occasions you can import them from a flat file (DME) or add them manually.  

Let's say we have a large enterprise application with lots of tables (1000+). Most of tables can be empty on the target environment. They contain irrelevant logging information or store other data that is not required for testing. Some tables contain important application data or contain master data like currencies, gender, types, products, etc. Those tables must be completely present in the target database. Other tables contain process data, like contracts, orders, addresses, customers and transactions. If we would select some of the customers, than we have to make sure that all data that belongs to that customer is selected accordingly. This is exactly what DATPROF Subset solves!

When you want to see details of a table, just right click on the table and click Properties... A popup window will open with the columns, foreign keys, dependencies and additional filters of the selected table. 

If you want to see graphical representation of your data model go to the upper menu bar and click Visualize → Data model. Make sure that that all the classifications are checked and click Visualize to generate a graph of your data model. Very large data models can take a while to render. 

If you got stuck, just take a look at the following video.