New column for specific value in numpy array

Question

Iam trying to prepare my data for machine learning. Iam recording values via the api of a simulationtool (CoppeliaSim).

My array is looking something like

x = ["start", 1, 3, 2, "start", 2, 4, "start", 3, 1, 2, 4, 5]

and i have to convert it to

x = [["start", 1, 3, 2], ["start", 2, 4], ["start", 3, 1, 2, 4, 5]]

So everytime i encounter "start" every entry after it should be a new column and with the next start a new row should start.

How can i manage this? Iam using python 3.7 but maybe there has to be a 2.7 version too.

Thanks!

Answer 1

You cannot have an array of multiple dtypes and non-rectangular. Your best bet is to use lists:

from more_itertools import split_before

x = list(split_before(x,lambda i:i=="start"))

output:

[['start', 1, 3, 2], ['start', 2, 4], ['start', 3, 1, 2, 4, 5]]