Search code examples
pandasdataframesktime

Dataframe to multiIndex for sktime format

I have a multivariate time series data which is in this format(pd.Dataframe with index on Time),

Existing format

I am trying to use sktime, which requires the data to be in multi index format. On the above if i want to use a rolling window of 3 on above data. It requires it in this format. Here pd.Dataframe has multi-index on (instance,time)

Desired format

I was thinking if it is possible to transform it to new format.

Solution

  • Edit here's a more straightforward and probably faster solution using row indexing

    df = pd.DataFrame({
    'time':range(5),
    'a':[f'a{i}' for i in range(5)],
    'b':[f'b{i}' for i in range(5)],
})

w = 3
w_starts = range(0,len(df)-(w-1)) #start positions of each window

#iterate through the overlapping windows to create 'instance' col and concat
roll_df = pd.concat(
    df[s:s+w].assign(instance=i) for (i,s) in enumerate(w_starts)
).set_index(['instance','time'])

print(roll_df)

    Output

                    a   b
instance time        
0        0     a0  b0
         1     a1  b1
         2     a2  b2
1        1     a1  b1
         2     a2  b2
         3     a3  b3
2        2     a2  b2
         3     a3  b3
         4     a4  b4