In Robot Framework, how to perform data-driven testing by creating a separate test case for each line of data in a text file?

Question

In Robot Framework, we can use Test Template to perform data-driven testing. However, in this method, the number of test cases are fixed. We are not able to add new test cas

Answer 1

I feel below one is the better approach.

Create two libraries one for reading the csv data and another for getting the number of lines like below.

csvLibrary.py

 1 import csv
 2 class csvLibrary(object):
 3
 4     def read_csv_file(self, filename):
 5         '''This creates a keyword named "Read CSV File"
 6
 7         This keyword takes one argument, which is a path to a .csv file. It
 8         returns a list of rows, with each row being a list of the data in
 9         each column.
 10         '''
 11         data = []
 12         with open(filename, 'rb') as csvfile:
 13             reader = csv.reader(csvfile)
 14             for row in reader:
 15                 data.append(row)
 16         return data

csvLibraryNoOfRows.py

1 import csv
2 class csvLibraryNoOfRows(object):
3
4     def csv_length(self, filename):
5         '''This creates a keyword named "CSV Length"
6
7         This keyword takes one argument, which is a path to a .csv file. It
8         returns a list of rows, with each row being a list of the data in
9         each column.
10         '''
11         length=0
12         with open(filename, 'rb') as csvfile:
13             reader = csv.reader(csvfile)
14             for row in reader:
15                 length+=1
16         return length

Include these two libraries in your test files. Using length, lets say "N" you can get the row data/cell data with the help of :FOR ${index} IN RANGE ${csvlength}

Example code is below.

Library        csvLibrary.py
Library        csvLibraryNoOfRows.py

*** Test Cases ***

Reading a csv file
${csvdata}=    read csv file   sample.csv
${csvlength}=   csv length  sample.csv
:FOR    ${index}    IN RANGE    ${csvlength}
    \   log ${csvdata[${index}]}

Answer 2

There is no way to directly do what you want. What you could do instead is write a script that reads your data file and auto-generates a robot test suite based on that data. The script that you use to run your tests could first run this other script to create the test file before running it.

You could also create the test suite via a suite setup, though I don't think I recommend doing it that way, as I don't think there's anything to gain and it makes your suite more complex.

Answer 3

I am not sure if you are really looking to create separate Test Cases (ie. with separate PASS / FAIL status etc in the output report), or simply looking to repeat a sequence of test steps using a set of data?

If the latter, you can easily read in lines from external files using the OperatingSystem library, parse the contents of the file using the String library, then repeatedly call a user keyword with the contents of each line.

| *** Settings ***
| Library | OperatingSystem | WITH NAME | os |
| Library | String | WITH NAME | str |

| *** Test Cases *** |
| Read Data From File |
| | ${fileContents}= | os.Get File | data.txt |
| | ${rows}= | str.Split To Lines | ${fileContents} |
| | :FOR | ${row} | IN | @{rows} |
| |      | ${cols}= | str.Split String | ${row} | , |
| |      | My Test Keyword | @{cols} |

| *** Keywords *** |
| My Test Keyword |
| | [Arguments] | @{fields} |
| | Log Many | ${fields} |

The first failure of My Test Keyword would normally fail the entire Read Data From File test case. If you wanted to run as many as possible, and then collate the results, use the Run Keyword And Ignore Error keyword from the BuiltIn library.