Skip to main content

A flexible utility for flattening and unflattening dict-like objects in Python.

Project description

https://img.shields.io/travis/ianlini/flatten-dict/master.svg https://img.shields.io/pypi/v/flatten-dict.svg https://img.shields.io/pypi/l/flatten-dict.svg https://img.shields.io/github/stars/ianlini/flatten-dict.svg?style=social

A flexible utility for flattening and unflattening dict-like objects in Python.

Introduction

This package provides a function flatten() for flattening dict-like objects in Python 2.7 and 3.5~3.8. It also provides some key joining methods (reducer), and you can choose the reducer you want or even implement your own reducer. You can also invert the resulting flat dict using unflatten().

Installation

pip install flatten-dict

Documentation

Flatten

def flatten(d, reducer='tuple', inverse=False, enumerate_types=(), keep_empty_types=()):
    """Flatten `Mapping` object.

    Parameters
    ----------
    d : dict-like object
        The dict that will be flattened.
    reducer : {'tuple', 'path', 'underscore', 'dot', Callable}
        The key joining method. If a `Callable` is given, the `Callable` will be
        used to reduce.
        'tuple': The resulting key will be tuple of the original keys.
        'path': Use `os.path.join` to join keys.
        'underscore': Use underscores to join keys.
        'dot': Use dots to join keys.
    inverse : bool
        Whether you want invert the resulting key and value.
    enumerate_types : Sequence[type]
        Flatten these types using `enumerate`.
        For example, if we set `enumerate_types` to ``(list,)``,
        `list` indices become keys: ``{'a': ['b', 'c']}`` -> ``{('a', 0): 'b', ('a', 1): 'c'}``.
    keep_empty_types : Sequence[type]
        By default, ``flatten({1: 2, 3: {}})`` will give you ``{(1,): 2}``, that is, the key ``3``
        will disappear.
        This is also applied for the types in `enumerate_types`, that is,
        ``flatten({1: 2, 3: []}, enumerate_types=(list,))`` will give you ``{(1,): 2}``.
        If you want to keep those empty values, you can specify the types in `keep_empty_types`:

        >>> flatten({1: 2, 3: {}}, keep_empty_types=(dict,))
        {(1,): 2, (3,): {}}

    Returns
    -------
    flat_dict : dict
    """

Examples

>>> from flatten_dict import flatten
>>> from pprint import pprint
>>> normal_dict = {
...     'a': '0',
...     'b': {
...         'a': '1.0',
...         'b': '1.1',
...     },
...     'c': {
...         'a': '2.0',
...         'b': {
...             'a': '2.1.0',
...             'b': '2.1.1',
...         },
...     },
... }
>>> pprint(flatten(normal_dict))
{('a',): '0',
 ('b', 'a'): '1.0',
 ('b', 'b'): '1.1',
 ('c', 'a'): '2.0',
 ('c', 'b', 'a'): '2.1.0',
 ('c', 'b', 'b'): '2.1.1'}
>>> pprint(flatten(normal_dict, reducer='path'))
{'a': '0',
 'b/a': '1.0',
 'b/b': '1.1',
 'c/a': '2.0',
 'c/b/a': '2.1.0',
 'c/b/b': '2.1.1'}
>>> pprint(flatten(normal_dict, reducer='path', inverse=True))
{'0': 'a',
 '1.0': 'b/a',
 '1.1': 'b/b',
 '2.0': 'c/a',
 '2.1.0': 'c/b/a',
 '2.1.1': 'c/b/b'}

The reducer parameter supports 'tuple', 'path', 'underscore', 'dot' and Callable. We can customize the reducer using a function:

>>> def underscore_reducer(k1, k2):
...     if k1 is None:
...         return k2
...     else:
...         return k1 + "_" + k2
...
>>> pprint(flatten(normal_dict, reducer=underscore_reducer))
{'a': '0',
 'b_a': '1.0',
 'b_b': '1.1',
 'c_a': '2.0',
 'c_b_a': '2.1.0',
 'c_b_b': '2.1.1'}

There is also a factory function make_reducer() to help you create customized reducer. The function currently only supports customized delimiter:

>>> from flatten_dict.reducer import make_reducer
>>> pprint(flatten(normal_dict, reducer=make_reducer(delimiter='_')))
{'a': '0',
 'b_a': '1.0',
 'b_b': '1.1',
 'c_a': '2.0',
 'c_b_a': '2.1.0',
 'c_b_b': '2.1.1'}

If we have some iterable (e.g., list) in the dict, we will normally get this:

>>> flatten({'a': [1, 2, 3], 'b': 'c'})
{('a',): [1, 2, 3], ('b',): 'c'}

If we want to use its indices as keys, then we can use the parameter enumerate_types:

>>> flatten({'a': [1, 2, 3], 'b': 'c'}, enumerate_types=(list,))
{('a', 0): 1, ('a', 1): 2, ('a', 2): 3, ('b',): 'c'}

We can even flatten a list directly:

>>> flatten([1, 2, 3], enumerate_types=(list,))
{(0,): 1, (1,): 2, (2,): 3}

If there is an empty dict in the values, by default, it will disappear after flattened:

>>> flatten({1: 2, 3: {}})
{(1,): 2}

We can keep the empty dict in the result using keep_empty_types=(dict,):

>>> flatten({1: 2, 3: {}}, keep_empty_types=(dict,))
{(1,): 2, (3,): {}}

Unflatten

def unflatten(d, splitter='tuple', inverse=False):
    """Unflatten dict-like object.

    Parameters
    ----------
    d : dict-like object
        The dict that will be unflattened.
    splitter : {'tuple', 'path', 'underscore', 'dot', Callable}
        The key splitting method. If a Callable is given, the Callable will be
        used to split `d`.
        'tuple': Use each element in the tuple key as the key of the unflattened dict.
        'path': Use `pathlib.Path.parts` to split keys.
        'underscore': Use underscores to split keys.
        'dot': Use underscores to split keys.
    inverse : bool
        Whether you want to invert the key and value before flattening.

    Returns
    -------
    unflattened_dict : dict
    """

Examples

>>> from pprint import pprint
>>> from flatten_dict import unflatten
>>> flat_dict = {
...     ('a',): '0',
...     ('b', 'a'): '1.0',
...     ('b', 'b'): '1.1',
...     ('c', 'a'): '2.0',
...     ('c', 'b', 'a'): '2.1.0',
...     ('c', 'b', 'b'): '2.1.1',
... }
>>> pprint(unflatten(flat_dict))
{'a': '0',
 'b': {'a': '1.0', 'b': '1.1'},
 'c': {'a': '2.0', 'b': {'a': '2.1.0', 'b': '2.1.1'}}}
>>> flat_dict = {
...     'a': '0',
...     'b/a': '1.0',
...     'b/b': '1.1',
...     'c/a': '2.0',
...     'c/b/a': '2.1.0',
...     'c/b/b': '2.1.1',
... }
>>> pprint(unflatten(flat_dict, splitter='path'))
{'a': '0',
 'b': {'a': '1.0', 'b': '1.1'},
 'c': {'a': '2.0', 'b': {'a': '2.1.0', 'b': '2.1.1'}}}
>>> flat_dict = {
...     '0': 'a',
...     '1.0': 'b/a',
...     '1.1': 'b/b',
...     '2.0': 'c/a',
...     '2.1.0': 'c/b/a',
...     '2.1.1': 'c/b/b',
... }
>>> pprint(unflatten(flat_dict, splitter='path', inverse=True))
{'a': '0',
 'b': {'a': '1.0', 'b': '1.1'},
 'c': {'a': '2.0', 'b': {'a': '2.1.0', 'b': '2.1.1'}}}

The splitter parameter supports 'tuple', 'path', 'underscore', 'dot' and Callable. We can customize the reducer using a function:

>>> def underscore_splitter(flat_key):
...     return flat_key.split("_")
...
>>> flat_dict = {
...     'a': '0',
...     'b_a': '1.0',
...     'b_b': '1.1',
...     'c_a': '2.0',
...     'c_b_a': '2.1.0',
...     'c_b_b': '2.1.1',
... }
>>> pprint(unflatten(flat_dict, splitter=underscore_splitter))
{'a': '0',
 'b': {'a': '1.0', 'b': '1.1'},
 'c': {'a': '2.0', 'b': {'a': '2.1.0', 'b': '2.1.1'}}}

There is also a factory function make_splitter() to help you create customized splitter. The function currently only supports customized delimiter:

>>> from flatten_dict.splitter import make_splitter
>>> pprint(unflatten(flat_dict, splitter=make_splitter(delimiter='_')))
{'a': '0',
 'b': {'a': '1.0', 'b': '1.1'},
 'c': {'a': '2.0', 'b': {'a': '2.1.0', 'b': '2.1.1'}}}

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for flatten-dict, version 0.3.0
Filename, size File type Python version Upload date Hashes
Filename, size flatten_dict-0.3.0-py2.py3-none-any.whl (8.2 kB) File type Wheel Python version py2.py3 Upload date Hashes View
Filename, size flatten-dict-0.3.0.tar.gz (8.9 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page