Does anyone know how use a value from keyword test A and use it in test B, also when updated in A?

Question

&nbsp;Hi TestCompleters!&nbsp;I have recorded two separate keyword tests, let's call them test A and test B.What I would like to do is to use a value created by an Object Checkpoint - generated in test A in test B and preferably in a way so that if the value changes in test A, then the new value is automatically transferred/updates in test B - is that possible?&nbsp;I reckon the solution somehow involves the creation of variables (I do not use script) and have tried to apply the method here:https://support.smartbear.com/screencasts/testcomplete/using-variables-in-keyword-tests/But come up against a brick wall when trying to locate the Objects Property Viewer - the video uses TestComplete version 7.50 and I use version 12.50 so I reckon that a lot has changed since then:&nbsp;- Does the "Objects Property Viewer" button still exist in TestComplete version 12.50? Or has it been renamed to "Object Spy"?&nbsp;Any suggestions are enthusiastically acccepted!Thank you in advance

tristaanogre · Answer

In Test A, use the operation "Set Variable Value" to create a variable on the Project level that contains the desired value you want to pass over to test B.&nbsp;&nbsp;
&nbsp;
Then, in Test B, whatever operation you are wanting to use the value in, there should be a parameter of some sort.&nbsp; That parameter has a "mode".&nbsp; Set that mode to "Variable" and then select the project variable.

tristaanogre · Answer

OK... totally different problem.&nbsp; My assumption earlier was that the asset description was something auto-generated by the application undertest and that you needed to track it through the test case.&nbsp; This is a problem where the application under test has built in checks to prevent creation of duplicate data records.&nbsp; So, each time through the test, you need to make sure that the asset description is unique.
&nbsp;
So... there are several different ways that this can be resolved, some more involved than others, some that may involve outside resourcing.&nbsp; And it all is going to depend somewhat on what is going to work best for you.&nbsp; Here are the high level descriptions.
&nbsp;
1) This is my preferred way, if possible.&nbsp; Your test environment should be treated as a "clean room" where you have control over all aspects of your environment, including the database against which your application is working. Before your test case is run, you would "reset" your environment to clear out any conditions, data, information, etc., that may be artifacts from previous tests and experimentations.&nbsp; In the past, I've done this by scripting an SQL restore from backup so that, when my tests start, they start with completely reset data and I don't have residual information.&nbsp; It doesn't need to be a 100% "clean" database, just reset to a known starting point.&nbsp; As you've discovered, test automation is very sensitive to initial conditions at the start of the test.&nbsp; If there is already data present when you create the test, the automation (unless you code it otherwise) will assume that the same data is present upon each test run.&nbsp; This is a more involved solution and requires you to have a level of access to your application under test, the data against which it runs, etc., that is beyond just a simple user but more on a sysadmin role.&nbsp; You can either do these things yourself or request others to do these resets for you on a scheduled basis.
&nbsp;
2) Second best option, IMO, is to create your data during your test using some sort of unique numbering scheme.&nbsp; A good way of doing this is to utilize the current date and time to create a unique number that you can then use in that initial text box when you enter the asset description.&nbsp; So, for example, if you do something like this as a "Code Expression" in the SetText parameter, this will give you a unique asset description every time your test is run
&nbsp;
'abc' + aqConvert.DateTimeToFormatStr(aqDateTime.Now(), '%Y%m%d%H%M%S')
At the time I'm writing this, this will enter into your description field the value&nbsp;abc20180523080725 being that I'm posting this on May 23rd, 2018, at 08:07:25 AM Eastern US Time.&nbsp; Everytime you run your test, a new number will be generated creating a unique asset description.&nbsp; The bonus of this is that you can see in your database a history of test assets created.&nbsp; Now, if you need to use that name at various points through your tests, use the "Set Variable Value" operation to set a variable equal to that code expression and then just use that variable in the appropriate parameters throughout your test case.
&nbsp;
3) Third option is to keep track of a counter.&nbsp; Several ways you can do this.&nbsp; Optimal is to keep the counter in some central location so that, no matter what machine you run against, you will always get the "next" value.&nbsp; Some sort of centralized INI file, XML file, or something that, when you start the test, you read the counter value into your test and append it to the end of your asset description using a code expression similar to what I describe in point 2 but, instead of that aqConvert blah blah blah, you would reference the variable value. After you read that counter into your test, increment it in that external source so that, next test run, you'll get a unique value.&nbsp; We do this for some of our tests where we store that counter in an SQL data table.&nbsp; If the external option won't work for you, you could do the same with a persistent project variable.&nbsp; The draw back to that is that, if you run the test on a second machine, you still have the risk of having duplicated numbers because the persistent variables are machine specific.
&nbsp;
4) This final option is the least desirable but it basically replaces the manual step you're doing by automating it.&nbsp; Add to your test code to detect if that warning message is present.&nbsp; If it is, have the code click the OK button to proceed.&nbsp; This is probably the simplest solution but, IMO, the least desirable because you are basically circumventing the application's built in checks against duplicate data and adding unnecessary steps to your test case.&nbsp; This can be developed as part of the recording process or you can go back in after the fact and add the steps in.
&nbsp;
So... these are the 4 options that I can see... there may be other ways but in my experience they would simply be variants of this.&nbsp; As mentioned, optimally, I'd prefer option 1 because then you have absolute control over your test environment and will make sure you're always testing the EXACT same thing each time rather than having residual artifacts left over from previous test runs.&nbsp; Option 2 is&nbsp; a VERY close second.&nbsp; 3 and 4 are OK, but they retain a higher risk of problems.
&nbsp;
Hope this helps give some more guidance.

tristaanogre · Answer

Just a note, if you need specific guidance how to implement something, feel free to ask and we'll be glad to help out.&nbsp; Something to keep in mind is that good, robust test automation goes beyond record/playback.&nbsp; It is, in truth, a type of software development so good development practices are key and you should learn how basic code practices work with setting variables, using if/then logic, writing for and while loops (all things available in Keyword tests so scripting isn't strictly necessary).&nbsp; But it would also be a good idea to learn a scripting language.&nbsp; There are plenty of online tutorials on how different script languages work.&nbsp; I still reference them when I need some additional help.&nbsp; For JavaScript, check out&nbsp;https://www.w3schools.com/js/default.asp

tristaanogre · Answer

I see you keep editing your post... but I'm not sure what the problem is.&nbsp; Object Spy/Object Property Viewer are not directly related to being able to use data from one Keyword Test in another (it's been quite a number of years since I've played with TestComplete 7 but I'm sure Object Spy is what you're looking for with regards to investigating an object).&nbsp; &nbsp;
&nbsp;
If the answer I gave was unhelpful, let me know what you still need and I'll try to help.

cmbdnne · Answer

&nbsp;Thanks a lot for your input so far.&nbsp;I've edited my post several times as I'm trying to describe my issue as clearly as possible (hasn't been very successful so far ;)&nbsp;1. I start recording a test, Test A. During this test an asset value is created and I checkpoint this value and keep on recording.&nbsp;2. I keep recording Test A and at some point need to&nbsp;insert the&nbsp;the value above in a text box - and&nbsp;I would like my test to do this automatically when I re-run Test A.This is because&nbsp;the value of the asset created in test step 1&nbsp;automatically goes from, e.g. 1 to 2 to 3 to 4 etc every time this test step is re-run&nbsp;3. Another issue is that during the creation of this asset, a description of the asset must be entered in another text box - this could e.g. be abc1000 or anything else.As this description DOES NOT automatically increases from abc1000 to abc1001 to abc1002 every time this test is re-run, I need a way to make this happen if possible?I've considered options such as a cleverly created variable, a test parameter or a data loop in which the description increments are loaded from an Excel sheet.&nbsp;As I admittedly am a TestComplete newbie, any test "recipes", as detailed as possible,&nbsp;are warmly welcomed.I can upload screen shots if my problem description still appears unclear.&nbsp;Have a nice weekend!&nbsp;Best regards,//Camilla&nbsp;

Forum Discussion

Does anyone know how use a value from keyword test A and use it in test B, also when updated in A?

7 Replies

Related Content

Chrome updated, now TC can't find browser

Does anyone have good examples of the user of Wait in keyword testing

Referencing the Keyword Test

Anyone tried Adobe Acrobat API?

Anyone Else Crashing TC v10?

Recent Discussions

Pulling in Dynamic Links from a web page

Multiple execution of a test plan with different values from an Excel table

Dynamic Object

Your thoughts about a REST API in TestComplete

Keyword Test - Click - use -1, -1 by default?