问题
I have a number of timeseries data in arrays and wish to extract values between given dates in the simplest way possible avoiding loops. Here's an example:
from numpy import *
from datetime import *
# datetime array
date_a=array([
datetime(2000,1,1),
datetime(2000,1,2),
datetime(2000,1,3),
datetime(2000,1,4),
datetime(2000,1,5),
])
# item array, indices corresponding to datetime array
item_a=array([1,2,3,4,5])
# extract items in a certain date range
# after a certain date, works fine
item_b=item_a[date_a >= (datetime(2000,1,3))] #Out: array([3, 4, 5])
# between dates ?
item_c=item_a[date_a >= (datetime(2000,1,3)) and date_a <= (datetime(2000,1,4))]
# returns: ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()
Is there a one-line solution to this? I have looked at numpy any() and all(), and also where(), without being able to find a solution. I appreciate any help and point-in-direction!
回答1:
If you want one-liner, then you can use
item_c=item_a[(date_a >= (datetime(2000,1,3))) * (date_a <= (datetime(2000,1,4)))]
回答2:
It's not clear to me why you are using the item_a variable. But to isolate the entries you want you can simply do:
>>> np.where(np.logical_and(date_a >= datetime(2000,1,3), date_a <= datetime(2000,1,4)))
(array([2, 3]),)
The resulting indexes are zero-based, so they correspond to the third and fourth element of your array.
EDIT: np is due to import numpy as np. Doing from numpy import * is in fact a very bad idea. You will overwrite built in functions such as sum and abs for example...
HTH!
回答3:
I think the following should work for you using List Comprehension
[item_a[i] for i in xrange(0,len(date_a)) if date_a[i] >= (datetime(2000,1,3)) and date_a[i] <= (datetime(2000,1,4))]
Select all items in item_a within range 0 <= i < length of date_a where datetime(2000,1,3) <= date_a[i] <= datetime(2000,1,4)
来源：https://stackoverflow.com/questions/8325089/extract-items-from-array-between-given-values-conditions