indexing on datetime object within a numpy array

Question

I have a numpy array containing date (as datetime object), prices and a category as an integer:

array([[datetime.date(2013, 6, 5), 11.42, 1],
   [datetime.date(2013, 6, 7), 63.97, 1],
   [datetime.date(2013, 6, 19), 3.92, 1],
   [datetime.date(2013, 6, 19), 16.25, 2],
   [datetime.date(2013, 6, 20), 11.0, 2],
   [datetime.date(2013, 6, 22), 32.72, 2],
   [datetime.date(2013, 6, 25), 16.6, 3],
   [datetime.date(2013, 6, 26), 2.95, 2],
   [datetime.date(2013, 7, 1), 6.27, 1],
   [datetime.date(2013, 7, 1), 2.95, 1]], dtype=object)

Suming up the prices if a category is met via index_cat=(array==2).any(axis=1), followed by by np.sum(array[index_cat][:,1]) is obvious.

What I want to achieve now is basically the same, but instead of selecting a category, I want to take the datetime object's month or month-year combination as a criterion.

So I thought a index_june=(array==datetime.dateime(month='06').any(axis=1) should do it, but, even with lots of searching, I could not find a way to do this.

So, how do I express this to match the datetime object considering the absence of wildcards?

Thanks a lot in advance!

dnalow · Accepted Answer · 2016-11-04 13:13:33Z

1

You can use numpy's vectorize function:

getmonth = np.vectorize(lambda d: getattr(d, "month"))
ind = getmonth(arr[:,0]) == 11

another way is list comprehension:

ind = np.array([a[0].month == 11 for a in arr])

edited Nov 4, 2016 at 13:13

answered Nov 4, 2016 at 13:08

dnalow

9844 silver badges14 bronze badges

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

2 Comments

nick Over a year ago

Great! Thank you for the quick and clear answer! I prefer the list comprehension way as it suits my python-knowledge better (for now)..

dnalow Over a year ago

It also seems to be slightly faster.

Collectives™ on Stack Overflow

indexing on datetime object within a numpy array

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related