Forum Discussion

venkireddy1991's avatar
venkireddy1991
Occasional Contributor
5 years ago

End to end test cases Generations

Hi Team,   I have a Data Driven Component test cases for multiple components... (Excel fies)   I would like to combbine these test cases and form End to End cases..   Looking to some one who ca...
  • richie's avatar
    richie
    5 years ago
    Hi venkireddy1991,

    Just adding to Rao's comment.
    Api testing allows you to do what can be termed point to point testing, so each point where a message is created/transformed/validated/consumed by database would include point to point tests.

    Example
    So if systemA generates a message (.json) that is published to an esb, the adaptor transforms to xml/soap which is then pushed to a listening, validated webservice (systemB) which then writes message to database.

    In the above example, i before now would create point to point tests to verify the adaptors transformations checking the following transformations: file (.json to soap), field (json "addressline1" to soap <line1> and <line2>) and value ("addressline1":"29 bennett road" to <line1>29</line1><line2>bennett road</line2>) and i'd create point to point tests to verify the soap message is consumed via the validated listening webservice and writes to database for schema valid requests and that schema invalid requests are rejected at the webservice validation.

    So in the above scenario i have types of point to point tests. The adaptor tests and the webservice tests and i'd ensure (depending on risk and prioritisation) that most if not all thr relevant permutations were exercised from my point to point tests.

    If you consider the above, there is a small test coverage gap as in my tests as im not using the systemA front end to generate the initial messages that are consumed via the esb.....i.e. the point to point tests do not verify thst systemA can successfully create the .json messages that are published to the esb, hence i would have a very small number of SIT/end to end tests that are created to cover off the coverage gap.

    These SIT/end to end tests would require using systemA's front end to generate the messages. So id either do this manually or quite often if systemA and systemB are both browser based, id automate my SIT/end to end tests using something like Selenium to record using systemA to create and publish message and then id use Selenium to record the message appearing within the front end of systemB.

    Do you mean end to end in this sort of scenario? ReadyAPI! is an API testing tool, and you can test APIs from a customer perspective using BDDs and gherkin syntax within ReadyAPI!, but ive never seen any value on writing your soapui tests using BDD style.

    Does this help?

    Can you clarify a bit more?

    Cheers,

    Rich
nmrao's avatar
nmrao
Champion Level 3
5 years ago
Some confusion with the summary "End to end test cases Generations". What do you mean by that?

ReadyAPI can be used to design you tests according to your business flows which may include calling multiple REST methods, passing the data between the steps. So, user have to design these tests.
  • richie's avatar
    richie
    Community Hero
    5 years ago
    Hi venkireddy1991,

    Just adding to Rao's comment.
    Api testing allows you to do what can be termed point to point testing, so each point where a message is created/transformed/validated/consumed by database would include point to point tests.

    Example
    So if systemA generates a message (.json) that is published to an esb, the adaptor transforms to xml/soap which is then pushed to a listening, validated webservice (systemB) which then writes message to database.

    In the above example, i before now would create point to point tests to verify the adaptors transformations checking the following transformations: file (.json to soap), field (json "addressline1" to soap <line1> and <line2>) and value ("addressline1":"29 bennett road" to <line1>29</line1><line2>bennett road</line2>) and i'd create point to point tests to verify the soap message is consumed via the validated listening webservice and writes to database for schema valid requests and that schema invalid requests are rejected at the webservice validation.

    So in the above scenario i have types of point to point tests. The adaptor tests and the webservice tests and i'd ensure (depending on risk and prioritisation) that most if not all thr relevant permutations were exercised from my point to point tests.

    If you consider the above, there is a small test coverage gap as in my tests as im not using the systemA front end to generate the initial messages that are consumed via the esb.....i.e. the point to point tests do not verify thst systemA can successfully create the .json messages that are published to the esb, hence i would have a very small number of SIT/end to end tests that are created to cover off the coverage gap.

    These SIT/end to end tests would require using systemA's front end to generate the messages. So id either do this manually or quite often if systemA and systemB are both browser based, id automate my SIT/end to end tests using something like Selenium to record using systemA to create and publish message and then id use Selenium to record the message appearing within the front end of systemB.

    Do you mean end to end in this sort of scenario? ReadyAPI! is an API testing tool, and you can test APIs from a customer perspective using BDDs and gherkin syntax within ReadyAPI!, but ive never seen any value on writing your soapui tests using BDD style.

    Does this help?

    Can you clarify a bit more?

    Cheers,

    Rich
    • sonya_m's avatar
      sonya_m
      SmartBear Alumni (Retired)
      5 years ago

      Richie, Rao, thank you for help! This is appreciated.

       

      venkireddy1991 Did the advice help? Let us know. 

      If you still require help, please fill us in on more details as the Community is asking. Thank you.