Sounds reasonable. What about duplicates?

I haven't given some thoughts about duplicates, but I guess we can have an "ignore_duplicates" argument as well.

This Lists should be equal keyword first compares length of two list and if it is not same it fails.
if we have ignore_order=True as argument, we can check first list's element in 2nd list and if it is available at any position, test should pass.

Pseudo code

If  ignore_order=True
    remove duplicates from lists
    Fail if  not same length
    if  element is not available in 2nd list, assert fail.
else  ##  order should be checked 
    compare item1 and item2

if ignore_order == True: 
    assert_true(item1 in list2, msg='Index %d%s' % (index,  " Not in second list"))
else:
    assert_equal(item1, item2, msg='Index %d%s' % (index, name))

As length of both lists are, same we need to check first list's element in 2nd list in case of ignore_order.

Hi @pekkaklarck,
Can you please review following _yield_list_diffs function.
need to import from collections import Counter

def lists_should_be_equal(list1, list2, ignore_order=False,msg=None, values=True,
                              names=None):
    ....
    diffs = list(_yield_list_diffs(list1, list2, names,ignore_order))
   ....

def _yield_list_diffs(list1, list2, names,ignore_order=False):
    **if ignore_order :
        assert_true(Counter(list1) == Counter(list2 ), msg='Lists  are  different')
    else:**
        for index, (item1, item2) in enumerate(zip(list1, list2)):         
            name = ' (%s)' % names[index] if index in names else ''
            try:
                assert_equal(item1, item2, msg='Index %d%s' % (index, name))
            except AssertionError as err:
                yield unic(err)
    ...

If we add these two lines before for loop in the existing given function will compare two UN-ordered lists.

Case 1:
['A', 'B', 'C', 'B']
['A', 'D', 'B', 'B', 'E']
lists_should_be_equal(my_list1,my_list2,ignore_order=True)
AssertionError: Lengths are different: 4 != 5

Case 2:
['A', 'B', 'C', 'B']
['A', 'D', 'B', 'B']
lists_should_be_equal(my_list1,my_list2,ignore_order=True)
AssertionError: Lists are not matching

Case 3:
['A', 'B', 'C', 'B']
['C', 'A', 'B', 'B']

PASS

Adding more tests and verifying it

• Duplicates not removed and compares correctly.
• No original list modified.

Lists Should Be Equal With Different Values Ignore_order :: FAIL L... | PASS |

Lists Should Be Equal With Different Values And Own Error Message ... | PASS |

Lists Should Be Equal With Different Values And Own And Default Er... | PASS |

Lists Should Be Equal With Named Indices As List Ignore_order :: F... | FAIL |
Lists are different:
Index 0 (a): 11 != 13
Index 2 (c): 13 != 11

Lists Should Be Equal With Named Indices As List With Too Few Valu... | PASS |

Lists Should Be Equal With Duplicate Value Ignore_order :: PASS li... | PASS |

Lists Should Be Equal With Duplicate Value Second Small List :: FA... | FAIL |
Lists are different

Lists Should Be Equal With Two Duplicate Value Ignore_order :: FAI... | PASS |

Lists Should Be Equal With Duplicate And Different Value Ignore_or... | FAIL |
Lists are different:
Index 2: A != E
Index 3: D != A

#2720

#2720

@bbpatel2001, I don't understand the usage of collections.Counter at all. I would assume just sorting lists before comparison would be the easiest solution. Are you interested to look at this again?

As Pekka notes, sorting the list contents and comparing those sorted lists is the easiest way to do this, and the way that I've handled this situation in other projects when I only care about list contents rather than order.

why not to use the "should be equal" keyword in buildin?

What about introducing Set concept into framework itself and into Collections with related Sets Should Be Equal , Remove From Set, Convert To Set, etc?

For now we have $scalar, @list and &dict, so why not ^set for example? But even without introducing ^set structure, keyword Sets Should Be Equal could also operate on lists, as they can be easily converted to sets.

I don't think sets are needed so often that they'd require a separate variable type. Convert To Set and possibly also other set relayed kws could be added to Collections, though. That requires a separate issue (and a PR).

May this one help?
In our company I just extend this library to custom RF library and then import it to compare.
#2717 (comment)

Hi @pekkaklarck could you please review PR #3469

is this PR #3469 RF 3.2 candidate?

This was implemented by PR #3469. Thanks a lot @bbpatel2001!