junekelly / python-datapath

Get and set values in deeply nested structures using JSONPath like syntax

python-datapath

Use JSONPath like strings to get and set values in a deeply nested data structure with graceful retrieval if the structure does not exist and autovivification when setting values.

Inspired by, but not completely compliant with the [JSONPath specification] (http://goessner.net/articles/JsonPath/)

Why should I use this library?

YOU REALLY SHOULDN'T - This is nowhere near ready to be used yet!

datapath lets you:

• Get values from deeply nested structures without caring if any of the intermediate keys exist
• Set values into data structures and have all intermediate values created for you
• Turn deeply nested structures into flat dicts and back

Compact paths

The datapath library supports compact paths which cut a small amount of the verboseness of the full JSONPath spec. A number of simplifying assumptions are made:

• All paths are local
• 'Naked' paths without a prefix are assumed to be dict keys
• Bracked paths are assumed to be on the local object

Examples:

JSONPath	datapath equivalents
$.a	a, .a, ["a"], ['a']
$.[7]	[7]
$.*	, .
$..a	..a
@..a	..a
@.a	a, .a, ["a"], ['a']

The JSONPath spec suggest '@.' for local anchoring and '$.' for root anchoring. Where anchoring is not specified or relevant datapath allows the omission of the leading identifiers for example:

JSONPath	datapath compact
$.a,	a
@.a	a
$.['a']	a
@.['a']	a
$.["a"]	a
@.["a"]	a

NOTE! - Currently datapath doesn't allow anchoring markers

Escaping

You can set a key with a dot, or any other reserved character by escaping it:

from datapath.crud import get_path
 
print get_path('a\\.b', {'a.b' : 5})  # 5!

Compliance levels

The datapath library does not support the following parts of JSONPath at the moment:

• List slices
• Selectors beyond '*' - no ability to do things like book[author='lily']
• Anchoring specifications: '@.', '$.'

Enough talking!

from datapath.crud import get_path, find_path, set_path

data = {
    'a': ['hello', 'world'],
    'b': 5
}
    
print get_path(data, 'a')     # ['hello', 'world']
print get_path(data, 'a[0]')  # 'hello'

find_path(data, '*.*')      # [['hello', 'world'], 5]

set_path({}, 'not_there[4].more', 'value')
# {'not_there': [None, None, None, {'more': 'value'}]}

Known issues

• This library doesn't work, and using it is a fools errand
• Attempting to set recursive paths (e.g. 'a..b') doesn't work
• Recursive paths in general are likely to have undefined behavior