| Article Index |
|---|
| IBM Rational Robot - Working with Datapools |
| Page 2 |
| Page 3 |
| Page 4 |
| Page 5 |
| Page 6 |
| Page 7 |
| Page 8 |
| Page 9 |
| Page 10 |
| Page 11 |
| Page 12 |
| All Pages |
·What is a datapool
·Planning and creating a datapool
·Data types
·Using datapools with session scripts
·Using datapools with GUI scripts
You should familiarize yourself with the concepts and procedures in this chapter before you begin to work with datapools.
NOTE: This chapter describes datapool access from scripts played back in a TestManager suite. For information about datapool access from GUI scripts
played back in Robot, see the online help for Rational Robot.
A datapool is a test dataset. It supplies data values to the variables in a script during script playback.
Datapools let you automatically pump test data to virtual testers under high-volume conditions that potentially involve hundreds of virtual testers performing thousands
of transactions. Typically, you use a datapool so that:
·Each virtual tester that runs the script can send realistic data (which can include unique data) to the server.
·A single virtual tester that performs the same transaction multiple times can send realistic data to the server in each transaction.
If you do not use a datapool during script playback, each virtual tester sends the same literal values to the server (the values that were captured when you recorded the script).
For example, suppose you record a session script that sends order number 53328 to a database server. If 100 virtual testers run this script, order number 53328 is sent to
the server 100 times. If you use a datapool, each virtual tester can send a different order number to the server.
Datapool Tools
You create and manage datapools with either Robot or TestManager, as follows:
This chapter describes how to perform all of these activities.
Rational TestManager User's Guide Managing Datapool Files
A datapool consists of two files:
Datapool values are stored in a text file with a .csv extension.
Datapool column names are stored in a specification(.spc) file. The Robot or TestManager software is always responsible for creating and maintaining this file. You should never edit this file directly. .csv and .spc files are stored in the Datapool directory of your Robot project. Unless you import a datapool, the Robot or TestManager software automatically creates and manages the .csv and .spc files based on instructions you provide through the user interface.
If you import a datapool, you are responsible for creating the .csv file and populating it with data. However, the Rational Test software is still responsible for creating and managing the .spc file for the imported datapool.
For information about importing datapools, see Rational TestManager User's Guide.
NOTE: TestManager automatically copies a .csv file to each Agent computer that uses it. If an Agent's .csv file becomes out-of-date, TestManager automatically updates it.
Datapool Cursor
The datapool cursor, or row-pointer, can be shared among all users who access the datapool, or it can be unique for each user.
Sharing a datapool cursor among all users allows for a coordinated test. Because each row in the datapool is unique, each user can share the same cursor and still send unique records to the database. Also, a shared cursor can be persistent across suite runs. For example, suppose the
last datapool row accessed in a suite run is row 100:
·If the cursor is persistent across suite runs, datapool row access resumes with row 101 the first time the datapool is accessed in a new suite run.
·If the cursor is not persistent, datapool row access resumes with row 1 the first time the datapool is accessed in a new suite run.
NOTE: GUI users can share a cursor when playback occurs in a TestManager site, but not when playback occurs in Robot.
For information about defining the scope of a cursor, see the description of the Cursor argument on page 10-16.
Row Access Order
Row access order is the sequence in which datapool rows are accessed at test runtime.
With GUI scripts, you can control the row access order of the datapool cursor through the sequence argument of the SQABasic SQADatapoolOpen command.
With session scripts, you can control row access order through the Access Order setting in the Robot Configure Datapool in Test Script dialog box. (See page 10-17.)
Datapool Limits
A datapool can have up to 150 columns if the Rational Test software automatically generates the data for the datapool, or 32,768 columns if you import the datapool
from a database or other source. Also, a datapool can have up to 2,147,483,647 rows.
What Kinds of Problems Does a Datapool Solve?
If you play back a script just once during a test run, that script probably does not need to access a datapool.
But often during a test run, and especially during performance testing, you need to run the same script multiple times - for example:
·During performance testing, you may run multiple instances of a script so that the script is executed many times simultaneously. (Remember, a virtual tester is one runtime instance of a script.)
·During functional and performance testing, you may often run multiple iterations of a script so that the script is executed many times consecutively (simulating a virtual tester performing the same task over and over). If the values used in each script instance and each script iteration are the same literal values - the values you provided during recording - you might encounter problems at test runtime.
Here are some examples of problems that datapools solve:
