如何在 python 的一个数据集中通过多个经度和纬度组合数据 arrays

Question

I made for loop of my data array of each long/lat point and I appended all result in one list as below:

out_list=[]
for i in ds.longitude.values:
    for j in ds.latitude.values:
        point = arr.sel(longitude=i,latitude=j)
        
        p_detrend = sm.tsa.tsatools.detrend(point, order=1,axis=0)
        out_list.append(p_detrend)

and my list as below: and you can see there are many arrays and each one has long/lat. How to combine all arrays in one dataset by longitude and latitude?

Answer 1

maybe you could use dictionary instead, something like:

out_list = []

for i in ds.longitude.values:
    for j in ds.latitude.values:
        point = arr.sel(longitude=i,latitude=j)
        p_detrend = sm.tsa.tsatools.detrend(point, order=1,axis=0)

        result = {
            'longitude': i, 
            'latitude': j,
            'detrend_arr': p_detrend  
        }
        out_list.append(result)

then you may use pandas to convert into pd.DataFrame for other manipulation;

Answer 2

What if you use xarray.merge() ?

I used the example from documentation and your description to create a fake data for different longitudes and latitudes, all stacked in a list.

Then, if I use xarray.merge() , it will automatically structure the data into ('time','longitude','latitude') xarray object.

import numpy as np
import pandas as pd
import xarray as xr

np.random.seed(123)

xr.set_options(display_style="html")

times = pd.date_range("2000-01-01", "2001-12-31", name="time")
annual_cycle = np.sin(2 * np.pi * (times.dayofyear.values / 365.25 - 0.28))

ds = []

for lon in range(-40,-30):
    for lat in range(0,10):
        base = 10 + 15 * annual_cycle.reshape(-1,1,1)
        tmin = base + 3 * np.random.randn(annual_cycle.size,1,1)
        tmax = base + 10 + 3 * np.random.randn(annual_cycle.size,1,1)

        ds.append(xr.Dataset(
            {
                "tmin": (("time", "latitude","longitude"), tmin),
                "tmax": (("time", "latitude","longitude"), tmax),
            },
            {"time": times,"latitude":[lat],"longitude":[lon]},
        ))
ds = xr.merge(ds)

First, ds will be like a list of different xarray objects for each latitude and longitude:

[<xarray.Dataset>
 Dimensions:    (latitude: 1, longitude: 1, time: 731)
 Coordinates:
   * time       (time) datetime64[ns] 2000-01-01 2000-01-02 ... 2001-12-31
   * latitude   (latitude) int64 0
   * longitude  (longitude) int64 -40
 Data variables:
     tmin       (time, latitude, longitude) float64 -4.646 -3.527 ... -3.939
     tmax       (time, latitude, longitude) float64 8.436 1.992 ... 3.366 1.605,
 <xarray.Dataset>
 Dimensions:    (latitude: 1, longitude: 1, time: 731)
 Coordinates:
   * time       (time) datetime64[ns] 2000-01-01 2000-01-02 ... 2001-12-31
   * latitude   (latitude) int64 1
   * longitude  (longitude) int64 -40
 Data variables:
     tmin       (time, latitude, longitude) float64 -2.966 -3.322 ... -7.731
     tmax       (time, latitude, longitude) float64 3.156 5.467 ... 11.85 7.26,
 <xarray.Dataset>
.
.
.
.
]

Then, after the merge, it will be a xarray.dataset just like this:

Please, let me know if that helps, or if I need to adapt in some way to work for your specific type of data.