Cannot parse datatable with examples

[sayurionishi]

Sample Scenario

Scenario Outline: User selects bookmark position
    When the user selects A in the bookmarks
    Then the position should be the following
      | x   | y   | z   |
      | <x> | <y> | <z> |

Examples:
  | x | y | z |
  | 1 | 2 | 3 |

Before, it could parse the values correctly, returning 1, 2, 3, but now it returns the placeholders as strings <x>, <y>, <z>.
This worked before the recent change in data tables.
Hope you could help check this issue, thank you

[jsa34]

Hello @sayurionishi

Could you please provide a full MVP?

This looks invalid gherkin syntax - there is no Feature, for example.

Could you also provide the step definitions used?

Also, you refer to datatables, but this has both examples table and datatable. Hopefully the step defs will clarify what you are trying to do: you seem to have both here, but none of the examples are actually being used in your feature file?

Try and use the bug reporting template provided when raising a bug:
Describe the bug
A clear and concise description of what the bug is.

To Reproduce
Write a Minimal, Reproducible Example to show the bug.

Expected behavior
A clear and concise description of what you expected to happen.

Additional context
Add any other context about the problem here.

Version

• pytest version:
• pytest-bdd version:
• OS:

[sayurionishi]

Good day, here it is @jsa34

Feature: Bookmarks Tests
Scenario Outline: User selects bookmark position
    When the user selects A in the bookmarks
    Then the position should be the following
      | x   | y   | z   |
      | <x> | <y> | <z> |

Examples:
  | x | y | z |
  | 1 | 2 | 3 |

@then(parsers.parse("the position should be the following"))
def then_position(datatable):
    expected_position = parse_table(datatable)
    current_position = {'x': 1, 'y': 2, 'z': 3}
    assert (
        current_position["x"] == expected_position["x"]
    ), f'actual: {current_position["x"]}, expected: {expected_position["x"]}'
    assert (
        current_position["y"] == expected_position["y"]
    ), f'actual: {current_position["y"]}, expected: {expected_position["y"]}'
    assert (
        current_position["z"] == expected_position["z"]
    ), f'actual: {current_position["z"]}, expected: {expected_position["z"]}'


def parse_table(data_table):
    headers = data_table[0]
    lst = [dict(zip(headers, row)) for row in data_table[1:]]

    if len(lst) == 1:
        return lst[0]
    return lst

Error: 1 was not equal to "<x>"
E       AssertionError: actual: 1, expected: <x>
E       assert '1' == '<x>'
E
E         - <x>
E         + 1

I have also edited the first comment as it does not display the datatable properly
pytest version: 8.1.1
pytest-bdd version: 8.0.0
OS: Windows

Thank you

[jsa34]

Thanks - that's fab.

I will take a look, but I haven't seen example table values used in a data table before...

[youtux]

Interesting use case, I wonder if other implementations treat data tables as part of the template, or not. Maybe to be raised with the Cucumber guys.

Sounds like a valid use case for me though, hopefully it won't be too hard to fix.

[jsa34]

Agreed - I have asked to get a conclusive idea of where example params should affect (for example, I know scenario names can use example params, too)

[jsa34]

Raised with Cucumber, and sought clarification.
The following need to be interpolated with examples table:

• Docstrings
• Datatables
• Step names (already done)
  I'll raise a PR to fix the first two