Converting pandas dataframe to 3D array as input to LSTM model

Question

Pandas dataframe:

id	feature1	feature2
1	2234	13
1	2235	12
2	3344	12
2	3364	14
2	3354	9

Problem:

There are more than 2 features in the data on which model needs to be trained. Each id can have multiple rows, so need to group by id to get the sequence data for each id. And then convert it to 3d numpy array as required by LSTM model

Already tried Tried solution here, and it works for single feature though I need it for multiple features.

Appreciate any help here.