连接 xarray.DataArray 并添加新维度答案

【问题标题】：concatenate xarray.DataArray and add new dimension连接 xarray.DataArray 并添加新维度
【发布时间】：2022-12-01 05:33:26
【问题描述】：

我有一个包含 1000 个键的字典，每个键有 23 个条目，每个条目都是一个 xarray.DataArray。每个条目如下所示：

<xarray.DataArray 'time' (time: 23)>
array(['1861-01-16T12:00:00.000000000', '1861-02-15T00:00:00.000000000',
       '1861-03-16T12:00:00.000000000', '1861-04-16T00:00:00.000000000',
       '1861-05-16T12:00:00.000000000', '1861-06-16T00:00:00.000000000',
       '1861-07-16T12:00:00.000000000', '1861-08-16T12:00:00.000000000',
       '1861-09-16T00:00:00.000000000', '1861-10-16T12:00:00.000000000',
       '1861-11-16T00:00:00.000000000', '1861-12-16T12:00:00.000000000',
       '1862-01-16T12:00:00.000000000', '1862-02-15T00:00:00.000000000',
       '1862-03-16T12:00:00.000000000', '1862-04-16T00:00:00.000000000',
       '1862-05-16T12:00:00.000000000', '1862-06-16T00:00:00.000000000',
       '1862-07-16T12:00:00.000000000', '1862-08-16T12:00:00.000000000',
       '1862-09-16T00:00:00.000000000', '1862-10-16T12:00:00.000000000',
       '1862-11-16T00:00:00.000000000'], dtype='datetime64[ns]')
Coordinates:
  * time     (time) datetime64[ns] 1861-02-15 ... 1862-12-16T12:00:00
    month    (time) int64 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11

我正在尝试连接所有这些条目并获得一个具有维度（1000,23）的新变量

看起来像

<xarray.DataArray '条目'，'时间'（条目：1000，时间：23）>

然后我希望能够将其写入 Netcdf 文件以便使用 .to_netcdf。

如果我做

tt=xr.concat(entry[0],entry[1])

我得到以下信息：

<xarray.DataArray 'time' (time: 23)>
array(['1861-01-16T12:00:00.000000000', '1861-02-15T00:00:00.000000000',
       '1861-03-16T12:00:00.000000000', '1861-04-16T00:00:00.000000000',
       '1861-05-16T12:00:00.000000000', '1861-06-16T00:00:00.000000000',
       '1861-07-16T12:00:00.000000000', '1861-08-16T12:00:00.000000000',
       '1861-09-16T00:00:00.000000000', '1861-10-16T12:00:00.000000000',
       '1861-11-16T00:00:00.000000000', '1861-12-16T12:00:00.000000000',
       '1862-01-16T12:00:00.000000000', '1862-02-15T00:00:00.000000000',
       '1862-03-16T12:00:00.000000000', '1862-04-16T00:00:00.000000000',
       '1862-05-16T12:00:00.000000000', '1862-06-16T00:00:00.000000000',
       '1862-07-16T12:00:00.000000000', '1862-08-16T12:00:00.000000000',
       '1862-09-16T00:00:00.000000000', '1862-10-16T12:00:00.000000000',
       '1862-11-16T00:00:00.000000000'], dtype='datetime64[ns]')
Coordinates:
  * time     (time) datetime64[ns] 1861-01-16T12:00:00 1861-02-15 ... 1862-11-16
    month    (time) int64 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11
<xarray.DataArray 'time' (time: 23)>
array(['1861-01-16T12:00:00.000000000', '1861-02-15T00:00:00.000000000',
       '1861-03-16T12:00:00.000000000', '1861-04-16T00:00:00.000000000',
       '1861-05-16T12:00:00.000000000', '1861-06-16T00:00:00.000000000',
       '1861-07-16T12:00:00.000000000', '1861-08-16T12:00:00.000000000',
       '1861-09-16T00:00:00.000000000', '1861-10-16T12:00:00.000000000',
       '1861-11-16T00:00:00.000000000', '1861-12-16T12:00:00.000000000',
       '1862-01-16T12:00:00.000000000', '1862-02-15T00:00:00.000000000',
       '1862-03-16T12:00:00.000000000', '1862-04-16T00:00:00.000000000',
       '1862-05-16T12:00:00.000000000', '1862-06-16T00:00:00.000000000',
       '1862-07-16T12:00:00.000000000', '1862-08-16T12:00:00.000000000',
       '1862-09-16T00:00:00.000000000', '1862-10-16T12:00:00.000000000',
       '1862-11-16T00:00:00.000000000'], dtype='datetime64[ns]')
Coordinates:
  * time     (time) datetime64[ns] 1861-02-15 ... 1862-12-16T12:00:00
    month    (time) int64 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11

那么我如何遍历它并将字典的条目转换为 xarran.DataArray 的新维度并获得 (1000,23) xarray.DataArray ？谢谢！

【问题讨论】：

标签： loops dictionary output netcdf python-xarray

【解决方案1】：

一些想法：

ds = xr.Dataset({'Entry_0': entry[0], 
'Entry_1': entry[1], 
'Entry_2': entry[2]
#... and so on
})

或者另一种做同样事情的方法是：

# create empty xarray dataset
ds = xr.Dataset({})
# add data arrays to xarray dataset
ds['Entry_0'] = entry[0]
ds['Entry_1'] = entry[1]
ds['Entry_2'] = entry[2]
# add so on

或者使用 for 循环以获得更简洁的代码：

# create empty list to store xarray data arrays
data_arrays = []
# for loop to add data arrays to list
for idx in range(len(entry)): 
    da = entry[idx]
    data_arrays.append(da)
# concatenate list of data arrays along new dimension into xarray dataset
ds = xarray.concat(data_arrays, dim='new_dim')

【讨论】：