Select xarray/pandas index based on specific months

Question

I have an xarray DataArray that I want to select the months April, May, June (similar to time.season=='JJA') for an entire time series.

Its structured like:

<xarray.DataArray 't2m' (time: 492, latitude: 81, longitude: 141)>

I have been previously selecting JJA by:

seasonal_data =temp_data.sel(time=temp_data['time.season']=='JJA')

I would like to do the same thing but with the months 'AMJ' instead. I can add any details that I might be missing.

Thanks

shoyer · Accepted Answer · 2016-10-26 21:21:55Z

29

The easiest way to select custom months is to use boolean masks, e.g.,

def is_amj(month):
    return (month >= 4) & (month <= 6)

seasonal_data = temp_data.sel(time=is_amj(temp_data['time.month']))

Note that you need to use the bitwise operators like & or | because Python's built-ins and and or don't work on vectors. Also, you need the parentheses because bitwise operators have higher precedence than comparisons.

answered Oct 26, 2016 at 21:21

shoyer

9,6831 gold badge41 silver badges57 bronze badges

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

4 Comments

Gabe Over a year ago

I had been trying to do this, but kept getting errors. This method worked super well. I assume that is how the time.season=='JJA' works in the underlying code?

shoyer Over a year ago

Take a look for yourself: github.com/pydata/xarray/blob/… We actually use a trick with integer division modulo 4, which is probably slightly faster but also less clear (and less generalizable).

Arindan Mandal Over a year ago

Hi @shoyer, Thanks for the example above. Its helped a lot. Additionally, I was wondering how do one can select, lets say, the Dec month from the previous year and a few months from the current year together. I tried the below function, but that gives an an empty array. (def is_djfma(month): return (month >= 12) & (month <= 4)) Do you have any quick suggestion for this? Thanks, Arindan.

ClimateUnboxed Over a year ago

@ArindanMandal wrapped months are easier to handle with the isin method.

ClimateUnboxed · Accepted Answer · 2024-05-16 08:28:33Z

2

Another way to do this is using the isin method:

# April, May and June:
data=temp_data.sel(time=temp_data.time.dt.month.isin([4,5,6]))

This makes handling "wrapped" seasons (e.g. DJFM) so much easier, in response to some of the comments under the accepted answer, e.g.:

# DJFM
data=temp_data.sel(time=temp_data.time.dt.month.isin([1,2,3,12]))

This is the recommended way to select months in the xarray documentation.

edited May 16, 2024 at 8:28

answered May 7, 2024 at 4:38

ClimateUnboxed

8,1866 gold badges48 silver badges100 bronze badges

Collectives™ on Stack Overflow

Select xarray/pandas index based on specific months

2 Answers 2

4 Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

4 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related