Skip to content

Reference

unique(values, *, key=lambda v: v)

Keep only the first occurrences while maintaining original order.

The equality check used for deduplication can be overridden using the key argument.

Examples:

>>> unique([1, 1, 5, 3, 3])
[1, 5, 3]
>>> unique([{'a': 1, 'b': 2}, {'a': 1, 'b': 3}], key=lambda v: v['a'])
[{'a': 1, 'b': 2}]
>>> unique([{'a': 1, 'b': 2}, {'a': 1, 'b': 3}], key=lambda v: v['b'])
[{'a': 1, 'b': 2}, {'a': 1, 'b': 3}]

Parameters:

Name Type Description Default
values Iterable[T]

An iterable containing your values

 required
key Callable[[T], Any]

Override the identity function of the equality check.

 lambda v: v

Returns:

Type Description
List[T]

A deduplicated list.
Source code in great_ai/utilities/unique.py 
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
def unique(values: Iterable[T], *, key: Callable[[T], Any] = lambda v: v) -> List[T]:
    """Keep only the first occurrences while maintaining original order.

    The equality check used for deduplication can be overridden using the `key` argument.

    Examples:
        >>> unique([1, 1, 5, 3, 3])
        [1, 5, 3]
        >>> unique([{'a': 1, 'b': 2}, {'a': 1, 'b': 3}], key=lambda v: v['a'])
        [{'a': 1, 'b': 2}]
        >>> unique([{'a': 1, 'b': 2}, {'a': 1, 'b': 3}], key=lambda v: v['b'])
        [{'a': 1, 'b': 2}, {'a': 1, 'b': 3}]

    Args:
        values: An iterable containing your values
        key: Override the identity function of the equality check.

    Returns:
        A deduplicated list.
    """

    key_values = {}
    for v in values:
        k = key(v)
        if k not in key_values:
            # dicts maintain insertion order: https://mail.python.org/pipermail/python-dev/2017-December/151283.html
            key_values[k] = v

    return list(key_values.values())
Last update: July 8, 2022