Exactly how do Visual Studio 2010 Data Sources behave when Access Method is "Random"?

Question

In Visual Studio 2010, if you bind a Data Source to a Web Performance Test you have the option of setting the Access Methods to "Random", defined as follows:

Move randomly through the rows in a table. This access method will loop through data in a table throughout the duration of a test.

We've been parsing this definition, but are not sure exactly what happens. Does it mean:

• Each time the source is accessed a row is chosen at random (i.e. you might get the same row in two tests in a row by chance); OR
• The source is first shuffled into a random order, and then the data source will "loop through" the shuffled data (i.e. every row is used once before any row is seen a second time); OR
• Something else?

Note we only have one agent, so repetition from that source is not a concern.

Thanks in advance.

Answer 1

Testing confirms that indeed the row is chosen entirely at random.

With a simple data source:

value
0
1
2
3

the order of values chosen in a test I just ran was:

3
3
3
1
1
2
3
...etc

For an actual "shuffle" implementation you'd need to write your own WebTestPlugin or WebTestRequestPlugin.