Python key functions for multi-field ordering in SQL ORDER BY fashion
Meant to be used with built-in sorted()
key function.
Supports also list.sort()
doing in-place sorting.
Implementation uses operator.itemgetter()
+ some internal helper classes to allow descending sorting order.
So far this is tested and used on lists of dictionaries. Adding support for named tuples and others would
be possible (using operator.attrgetter()
).
- SQL-like:
orderby('foo ASC, bar DESC')
- chained:
asc('foo').desc('bar')
usage - multiple fields at once:
asc('foo', 'bar')
orderby()
string syntax:
>>> from orderby import orderby
>>> import json
>>> files = [
... {'size': 1234, 'path': 'foo/bar.txt'},
... {'size': 0, 'path': '/dev/null'},
... {'size': 1234, 'path': 'foo/abc.bin'},
... ]
>>> print(json.dumps(sorted(files, key=orderby('size DESC, path')), indent=2))
[
{
"size": 1234,
"path": "foo/abc.bin"
},
{
"size": 1234,
"path": "foo/bar.txt"
},
{
"size": 0,
"path": "/dev/null"
}
]
Chained asc() and desc() usage:
>>> from orderby import asc, desc
>>> print(json.dumps(sorted(files, key=desc('size').asc('path')), indent=2))
[
{
"size": 1234,
"path": "foo/abc.bin"
},
{
"size": 1234,
"path": "foo/bar.txt"
},
{
"size": 0,
"path": "/dev/null"
}
]
In-place sorting of a list:
>>> files.sort(key=desc('path'))
>>> print(json.dumps(files, indent=2))
[
{
"size": 1234,
"path": "foo/bar.txt"
},
{
"size": 1234,
"path": "foo/abc.bin"
},
{
"size": 0,
"path": "/dev/null"
}
]
>>> files.sort(key=desc('size').asc('path'))
>>> print(json.dumps(files, indent=2))
[
{
"size": 1234,
"path": "foo/abc.bin"
},
{
"size": 1234,
"path": "foo/bar.txt"
},
{
"size": 0,
"path": "/dev/null"
}
]
Works also with SortedContainers:
>>> from sortedcontainers import SortedList
>>> from orderby import desc
>>> mylist = SortedList(key=desc('value'))
>>> mylist
SortedListWithKey([], key=<orderby.sorter.OrderBy object at 0x108f65978>, load=1000)
>>> mylist.add({'value': 13})
>>> mylist.add({'value': 2})
>>> mylist.add({'value': 1000})
>>> mylist
SortedListWithKey([{'value': 1000}, {'value': 13}, {'value': 2}], key=<orderby.sorter.OrderBy object at 0x108f65978>, load=1000)
To be explained here later...