I'm nesting 2 levels of default dicts. The inner dict contains a number of fields, and I would like to sort it by one of the values and delete the entries that correspond to the lowest values.
Here's a simplified code example:
from collections import defaultdict
sampleDict = defaultdict(lambda: defaultdict(lambda:defaultdict(lambda:str)))
sampleDict['keyA']['keyB']['size'] = 1000
sampleDict['keyA']['keyC']['size'] = 500
sampleDict['keyA']['keyD']['size'] = 750
sampleDict['keyA']['keyE']['size'] = 250
sampleDict['keyA']['keyB']['desc'] = 'some data'
sampleDict['keyA']['keyC']['desc'] = 'some more data'
sampleDict['keyA']['keyD']['desc'] = 'different data'
sampleDict['keyA']['keyE']['desc'] = 'other data'
In this case, I want to sort and identify that the highest size is ['keyA']['keyB'] and that the second highest is ['keyA']['keyD'] and then remove ['keyA']['keyC'] and ['keyA']['keyE'].
The reason it is nested is because I will then be looping through other entries in the outer dict.
Try this:
>>> import operator
>>> sorted(
... reduce(operator.add,
... [[(k, k1, sampleDict[k][k1]['size']) for k1 in v.keys()]
... for k,v in sampleDict.items()]
... ),
... key=lambda x: x[2], reverse=True)
[('keyA', 'keyB', 1000), ('keyA', 'keyD', 750), ('keyA', 'keyC', 500), ('keyA', 'keyE', 250)]
The reduce statement is used to turn the nested list [[a],[b,c],[d]] into [a,b,c].
The key argument for the sorted statement specifies to sort on the (zero inclusive) 2nd argument of (k,k1,val) ie, val.
The reverse argument orders the list in descending order.