slice second level datetime multindex dataframe

Question

I have a dataframe that looks like that:

I am trying to slice for a particular symbol and as from a date. doing

df.loc[symbol]

I obtain that dataframe:

So it goes from 2021-3-1 to 2023-1-31. I have a variable "date" = Timestamp('2021-03-01 00:00:00'), so the first date.

When I try to slice it like that:

df.loc[pd.IndexSlice[symbol, date:]]

I get the error:

KeyError: Timestamp('2021-03-01 00:00:00')

same for:

df.loc[pd.IndexSlice[symbol, :date]]

When I print

df.index[0]

('BOTTO/WETH', Timestamp('2023-01-31 00:00:00'))

to prove that the second level is indeed a timetamp.

Any idea how I could do please?

Answer 1

For me working well if added : for select all columns before last ]:

df = pd.DataFrame(

    np.random.randn(20, 1),

    columns=["A"],

    index=pd.MultiIndex.from_product(

        [["a", "b"], pd.date_range("20210228", periods=10, freq="D")]

    ),

)

---

symbol = 'a'
date = pd.Timestamp('2021-03-01 00:00:00')

df = df.loc[pd.IndexSlice[symbol, date:], :]
print (df)
                     A
a 2021-03-01  0.228366
  2021-03-02 -1.092949
  2021-03-03 -1.086836
  2021-03-04 -1.155692
  2021-03-05  0.382733
  2021-03-06 -1.653822
  2021-03-07  0.978806
  2021-03-08 -1.537005
  2021-03-09 -0.200136