我需要在一个numpy数组中找到值的位置，并使用它来引用另一个numpy数组的相同位置中的值

Question

I am writing code to analyze accelerometer data on a raspberry pi. The sensor data is output to a single txt file with columns separated by \\t. I imported the text file using numpy.loadtxt and unpacked it into separate arrays. I can perform things like trapz and cumtrapz on the arrays.

This data will be used in combination with another sensor that will output a specific time of an event. I want to take that time, find the closest logged time from my sensor and correspond it to values from the other arrays.

I tried using numpy.where with a specific time value that i knew was in the list and got an output of "(array([], dtype=int32),)"

Here is the code I ran. I'm sure I misused at least one thing. I am still very much a beginner in Python and coding in general...

import logging
import sys
import numpy as np
from scipy import integrate
x,y,z,t=np.loadtxt('a.txt', dtype={'names':['x','y','z','t'], 
'formats':['f4','f4','f4','f4']},unpack='true')
p = integrate.trapz(integrate.cumtrapz(x, t, initial=0), t)
ti = np.where(x==1.5670002)
print ti
print p

The full output from that is

(array([], dtype=int32),)
0.0114166

So I was searching x for a value from t. it is now outputting (array([101]),)

How would I print that corresponding number from another array?

Answer 1

Here is my solution:

import logging
import sys
import numpy as np
from scipy import integrate
x,y,z,t=np.loadtxt('a.txt', dtype={'names':['x','y','z','t'], 
'formats':['f4','f4','f4','f4']},unpack='true')
p = integrate.cumtrapz(integrate.cumtrapz(x, t, initial=0), t)
t0=input("What is reference time?")
ti = np.where(t>=t0)[0][0]
if t[ti]-t0 <= t0-t[ti-1]:
    t1 = ti
else:
    t1 = ti-1

print ('closest time was {0:0.4f}\ndisplacement at that time was {1:0.4f}' .format(t[t1],p[t1]))

output is

closest time was 1.1991
displacement at that time was 0.0100

Seems to be working. I will have to add error messages for when the reference time is outside of a usable range. Would love some constructive criticism though. Any commands that you think would work better/faster than what I have used?