dataframe to frozenset

Question

I want to translate a dataframe to frozensets and keep the dataframe columns within the frozenset.

Example

x=pd.DataFrame(data=dict(sample=["A","B","C"],lane=[1,1,2]))
>>> x
   lane sample
0     1      A
1     1      B
2     2      C

And I would like something as :

x2= {frozenset({("sample", "A"), ("lane", 1)}),
    frozenset({("sample", "B"), ("lane", 1)}),
    frozenset({("sample", "C"), ("lane", 2)})}

>>> x2
{frozenset({('sample', 'B'), ('lane', 1)}), frozenset({('sample', 'A'), ('lane', 1)}), frozenset({('lane', 2), ('sample', 'C')})}

I tried x.apply(frozenset,1) but it gives me that :

0    (1, A)
1    (1, B)
2    (C, 2)
dtype: object

Any help will be useful. Thank you

Answer 1

You can convert your dataframe to the records format you need with pd.DataFrame.to_dict :

x.to_dict('records')

# [{'sample': 'A', 'lane': 1}, 
#  {'sample': 'B', 'lane': 1}, 
#  {'sample': 'C', 'lane': 2}]

Since this results in a list, you can then map frozenset to the list like so:

# using abbreviation 'r' instead of 'records'
map(lambda y: frozenset(y.iteritems()), x.to_dict('r'))

# [frozenset([('sample', 'A'), ('lane', 1)]), 
#  frozenset([('sample', 'B'), ('lane', 1)]), 
#  frozenset([('sample', 'C'), ('lane', 2)])]

Or, using a set comprehension, if your output should be a set of frozensets:

{frozenset(y.iteritems()) for y in x.to_dict('records')}

# set([frozenset([('sample', 'C'), ('lane', 2)]),  
#      frozenset([('sample', 'B'), ('lane', 1)]), 
#      frozenset([('sample', 'A'), ('lane', 1)])])