How to slice Numpy datetime64 array

Question

I have a python zip with 2 arrays

zipped = zip(array1[], array2[])

Where array1 is of type numpy.datetime64[] adn array2 is a temperature

I want to make a time window in 1st array so i can have fixed array len (because i have other zipped arrays but they differ in array length)

this is what i have:

start = np.datetime64('2016-06-17T15:00')
stop = np.datetime64('2016-06-19T15:00')

index, temp = sensor_cal.get_arrays('ParsedData/parsed.csv')
print(index)
print(temp)
index2 = index[start:stop] /////////////This doesn't work
print(index2)

How can i define a time window like this.... My objective is to get same length arrays in the same time window (because they were previously frequency normalized) and then make a graph where xAxis is time and the various series correspond to the multiple temperature sensor arrays

My error:

['2016-06-17T13:23:59.000000000' '2016-06-17T13:24:59.000000000' '2016-06-17T13:25:59.000000000' ..., '2016-06-20T09:55:59.000000000' '2016-06-20T09:56:59.000000000' '2016-06-20T09:57:59.000000000'] [[
nan] [ nan] [ nan] ..., [ 25.54 ] [ 25.56333333] [ 25.59333333]] Traceback (most recent call last): File "main_cal.py", line 10, in index[start:stop] IndexError: failed to coerce slice entry of type numpy.datetime64 to integer

In index[start:stop], start and stop must be integers, indexes to positions in the array, e.g. index[2:5]. They are not values. You may need to review what indexing and slicing does in numpy (or for that matter Python lists). — hpaulj
– hpaulj, Commented Jun 22, 2016 at 17:38

Yarnspinner · Accepted Answer · 2016-06-23 14:10:57Z

3

You can use pandas indexing which is designed for this. 'series' is a 1d array with an index attached. With reference to Wes McKinney's Python for Data Analysis:

import pandas as pd
temp = np.random.randn(366)
time_series = pd.Series(temp,index=np.arange(np.datetime64('2015-12-19'),np.datetime64('2016-12-19')))

start = np.datetime64('2016-01-17T15:00')
stop = np.datetime64('2016-06-19T15:00')
time_series[start:stop]

Output:

2016-01-18   -0.690170
2016-01-19   -0.638598
2016-01-20    0.231680
2016-01-21   -0.202787
2016-01-22   -1.333620
2016-01-23    1.525161
2016-01-24   -0.908140
2016-01-25    0.493663
2016-01-26   -1.768979
2016-01-27    0.147327
...

edited Jun 23, 2016 at 14:10

answered Jun 23, 2016 at 13:49

Yarnspinner

8925 silver badges8 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

How to slice Numpy datetime64 array

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related