Search code examples
pythonpandasdatetimetime-seriesmulti-index

What is the proper way of selecting date ranges in Pandas multi-indexes?

What is the proper way of selecting date ranges in Pandas multi-indexes?

I've got a multi-index dataframe, that looks like the following:

enter image description here

If I wish to select a particular day, it's trivial using xs:

data.xs('2011-11-11', level='Date').head()

However if I wish to select a date range, I cannot. All the the following give me an Invalid Syntax error: 

data.xs('2011-10-10':'2011-11-11', level='Date').head()
data.xs(['2011-10-10':'2011-11-11'], level='Date').head()

Note #1: I am looking for a way to use elegant Pandas functionality. Naturally it's easy enough to hack around the problem using 4 or 5 lines of code, the question is about what the "right way" is.

Note #2: I've seen this answer, but that didn't cover this case.

Solution

  • Using data from previous question:

    d = {'Col1': {(Timestamp('2015-05-14 00:00:00'), '10'): 81.370003,
  (Timestamp('2015-05-14 00:00:00'), '11'): 80.41999799999999,
  (Timestamp('2015-05-14 00:00:00'), 'C3'): 80.879997,
  (Timestamp('2015-05-19 00:00:00'), '3'): 80.629997,
  (Timestamp('2015-05-19 00:00:00'), 'S9'): 80.550003,
  (Timestamp('2015-05-21 00:00:00'), '19'): 80.480003,
  (Timestamp('2015-05-22 00:00:00'), 'C3'): 80.540001},
 'Col2': {(Timestamp('2015-05-14 00:00:00'), '10'): 6.11282,
  (Timestamp('2015-05-14 00:00:00'), '11'): 6.0338,
  (Timestamp('2015-05-14 00:00:00'), 'C3'): 6.00746,
  (Timestamp('2015-05-19 00:00:00'), '3'): 6.10465,
  (Timestamp('2015-05-19 00:00:00'), 'S9'): 6.1437,
  (Timestamp('2015-05-21 00:00:00'), '19'): 6.16096,
  (Timestamp('2015-05-22 00:00:00'), 'C3'): 6.1391599999999995},
 'Col3': {(Timestamp('2015-05-14 00:00:00'), '10'): 39.753,
  (Timestamp('2015-05-14 00:00:00'), '11'): 39.289,
  (Timestamp('2015-05-14 00:00:00'), 'C3'): 41.248999999999995,
  (Timestamp('2015-05-19 00:00:00'), '3'): 41.047,
  (Timestamp('2015-05-19 00:00:00'), 'S9'): 41.636,
  (Timestamp('2015-05-21 00:00:00'), '19'): 42.137,
  (Timestamp('2015-05-22 00:00:00'), 'C3'): 42.178999999999995},
 'Col4': {(Timestamp('2015-05-14 00:00:00'), '10'): 44.950001,
  (Timestamp('2015-05-14 00:00:00'), '11'): 44.75,
  (Timestamp('2015-05-14 00:00:00'), 'C3'): 44.360001000000004,
  (Timestamp('2015-05-19 00:00:00'), '3'): 40.98,
  (Timestamp('2015-05-19 00:00:00'), 'S9'): 42.790001000000004,
  (Timestamp('2015-05-21 00:00:00'), '19'): 43.68,
  (Timestamp('2015-05-22 00:00:00'), 'C3'): 43.490002000000004}}

df = pd.Dataframe(d)

    Then you can use partial string indexing to select a range of dates:

    df.loc['2015-05-14':'2015-05-19']

    Output:

                        Col1     Col2    Col3       Col4
2015-05-14 10  81.370003  6.11282  39.753  44.950001
           11  80.419998  6.03380  39.289  44.750000
           C3  80.879997  6.00746  41.249  44.360001
2015-05-19 3   80.629997  6.10465  41.047  40.980000
           S9  80.550003  6.14370  41.636  42.790001