I have a file f which holds N (unknown) events. Each event carries an (unknown and different for each event, call it i, j etc) amount of reconstructed tracks. Then, each track has properties like energy E and likelihood lik. So,
>>> print(f.events.tracks.lik)
[[lik1, lik2, ..., likX], [lik1, lik2, ..., likj], ..., [lik1, lik2, ..., likz]]
prints an array holding N subarrays (1 per event), each presenting the lik for all its tracks.
GOAL: call f.events.tracks[:, Inds].E to get the energies for the tracks with max likelihood.
Minimal code example
>>>import numpy as np
>>>lik = np.random.randint(low=0, high=100, size=50).reshape(5, 10)
>>>print(lik)
[[ 3 49 27 3 80 59 96 99 84 34]
[88 62 61 83 90 9 62 30 92 80]
[ 5 21 69 40 2 40 13 63 42 46]
[ 0 55 71 67 63 49 29 7 21 7]
[40 7 68 46 95 34 74 88 79 15]]
>>>energy = np.random.randint(low=100, high=2000, size=50).reshape(5, 10)
>>>print(energy)
[[1324 1812 917 553 185 743 358 877 1041 905]
[1407 663 359 383 339 1403 1511 1964 1797 1096]
[ 315 1431 565 786 544 1370 919 1617 1442 925]
[1710 698 246 1631 1374 1844 595 465 908 953]
[ 305 384 668 952 458 793 303 153 661 791]]
>>> Inds = np.argmax(lik, axis=1)
>>> print(Inds)
[2 1 8 6 7]
PROBLEM:
>>> # call energy[Inds] to get
# [917, 663, 1442, 1844, 153]
What is the correct way of accessing these energies?
You can select the values indexed by Inds for each line using a 2D indexing with a temporary array containing [0,1,2,...] (generated using np.arange).
Here is an example:
energy[np.arange(len(Inds)), Inds]