Parameterized Testing From CSV rows

Table Of Contents

• Introduction
• Example Input Data
• Test Scenario
• Type Casting(optional)
• Conclusion

Introduction

During the user journey, if the scenario to be validated has a number of similar set of input, then we can arrange them by comma separated values called CSV and fire the tests.

Example Input Data

When the source is a simple list of CSV rows as below:

//user - {0}   //name - {1}     //city - {2}         //id - {3}
"octocat,      The Octocat,     San Francisco,       583231",
"foo,          Foo Is My Name,  Bar City,            112233"

It looks like below in a table-format:

user	name	city	id
octocat	The Octocat	San Francisco	583231
foo	Foo Is My Name	Bar City	112233

Test Scenario

In Zerocode we arrange the test input like below:

    "parameterized": {
        "csvSource":[
              "octocat,      The Octocat,     San Francisco,       583231",
              "foo,          Foo Is My Name,  Bar City,            112233"
        ]
    }

The full test scenario looks like below and runs twice for two(2) for the CSV rows i.e. once for each row.

{
    "scenarioName": "Fetch and assert GitHub userIds by their userNames",
    "steps": [
        {
            "name": "get_user_details",
            "url": "/users/${0}",
            "operation": "GET",
            "request": {
            },
            "assertions": {
                "status": 200,
                "body": {
                    "login" : "${0}",
                    "type" : "User",
                    "name" : "${1}",
                    "location" : "${2}",
                    "id" : "$EQ.${3}"
                }
            }
        }
    ],
    "parameterized": {
        "csvSource":[
              "octocat,      The Octocat,     San Francisco,       583231",
              "foo,          Foo Is My Name,  Bar City,            112233"
        ]
    }
}

Type Casting(optional)

While driving the tests from the set of input data, all input field are String by default. But if we want to use any of the fields as (int) or (decimal) or (boolean) etc, then we can cast them to the needed type as below.

{
    "scenarioName": "Parameterized and type casted example - GET API",
    "steps": [
        {
            "name": "get_user_details",
            "url": "/users/${0}",
            "operation": "GET",
            "request": {
            },
            "assertions": {
                "status": 200,
                "body": {
                    "login" : "${0}",
                    "type" : "User",
                    "name" : "${1}",
                    "location" : "${2}",
                    "id" : "(int)${3}",
                    "site_admin" : "(boolean)${4}"
                }
            }
        }
    ],
    "parameterized": {
        "csvSource":[
              "octocat,      The Octocat,     San Francisco,       583231, false",
              "foo,          Foo Is My Name,  Bar City,            112233, true"
        ]
    }
}

Conclusion

We learned here how we can do parameterized testing and how(optionally) we can do type-casting of a field to the desired type.

The examples are available in the HelloWorld section of the README file. We can clone this repo and try at home!