NetCDF时间序列切片与python3

Question

我试图从NetCDF文件绘制一周的时间序列数据，但遇到了一些问题。

我正在使用以下软件包：

import netCDF4
from matplotlib import pyplot as plt
import numpy as np
import xarray as xr
import dask

首先，我导入两个.nc文件：

^{pr2}$

然后我使用xarray选择时间和网格位置：

dsloc1 = ds1.sel(time='2008-02-10',longitude=2.2,latitude=48.7,method='nearest')
dsloc2 = ds2.sel(time='2008-02-10',longitude=2.2,latitude=48.7,method='nearest')

然后我画出两个时间序列：

dsloc1['v100'].plot.line('b-',figsize=(15,10))
dsloc2['v100'].plot.line('y-')

产生了我所期待的：

但是，当我试图选择一系列的日期时，我得到一些错误。。。

dsloc1 = ds1.sel(time=slice('2008-03-01','2008-03-07'),longitude=2.2,latitude=48.7,method='nearest')
dsloc2 = ds2.sel(time=slice('2008-03-01','2008-03-07'),longitude=2.2,latitude=48.7,method='nearest')

我肯定这可能是个语法问题，但我花了太多的时间来解决它。。如有任何建议，不胜感激！

[编辑]回溯如下：

---------------------------------------------------------------------------
NotImplementedError                       Traceback (most recent call last)
<ipython-input-308-c3385fa732ab> in <module>()
      1 # select time and grid location (Feb 10th, 2008, near Paris)
----> 2 dsloc1 = ds1.sel(time=slice('2008-03-01','2008-03-07'),longitude=2.2,latitude=48.7,method='nearest')
      3 dsloc2 = ds2.sel(time=slice('2008-03-01','2008-03-07'),longitude=2.2,latitude=48.7,method='nearest')

/usr/local/lib/python3.6/site-packages/xarray/core/dataset.py in sel(self, indexers, method, tolerance, drop, **indexers_kwargs)
   1507         indexers = either_dict_or_kwargs(indexers, indexers_kwargs, 'sel')
   1508         pos_indexers, new_indexes = remap_label_indexers(
-> 1509             self, indexers=indexers, method=method, tolerance=tolerance)
   1510         result = self.isel(indexers=pos_indexers, drop=drop)
   1511         return result._replace_indexes(new_indexes)

/usr/local/lib/python3.6/site-packages/xarray/core/coordinates.py in remap_label_indexers(obj, indexers, method, tolerance, **indexers_kwargs)
    353 
    354     pos_indexers, new_indexes = indexing.remap_label_indexers(
--> 355         obj, v_indexers, method=method, tolerance=tolerance
    356     )
    357     # attach indexer's coordinate to pos_indexers

/usr/local/lib/python3.6/site-packages/xarray/core/indexing.py in remap_label_indexers(data_obj, indexers, method, tolerance)
    248         else:
    249             idxr, new_idx = convert_label_indexer(index, label,
--> 250                                                   dim, method, tolerance)
    251             pos_indexers[dim] = idxr
    252             if new_idx is not None:

/usr/local/lib/python3.6/site-packages/xarray/core/indexing.py in convert_label_indexer(index, label, index_name, method, tolerance)
    132         if method is not None or tolerance is not None:
    133             raise NotImplementedError(
--> 134                 'cannot use ``method`` argument if any indexers are '
    135                 'slice objects')
    136         indexer = index.slice_indexer(_sanitize_slice_element(label.start),

NotImplementedError: cannot use ``method`` argument if any indexers are slice objects

Answer 1

似乎不支持使用时间片与method='nearest'结合使用的sel：

cannot use method argument if any indexers are slice objects

这在某种程度上是有意义的，因为为一个切片选择nearest似乎有点奇怪。在

您可以通过两个步骤执行sel来解决这个问题，即首先选择时间片，然后从该时间片中选择一个位置（或相反）。我不确定这是否是最好的解决办法，但至少它能起作用。在

一些ERA5数据的快速示例：

import xarray as xr

ds1 = xr.open_dataset('20160502_cabauw_model_fc.nc')

# Works:
dsloc1 = ds1.sel(time='2016-05-02 10:00', longitude=4.9, latitude=51.2, method='nearest')

# Doesn't work:
#dsloc2 = ds1.sel(time=slice('2016-05-02 10:00', '2016-05-02 12:00'), longitude=4.9, latitude=51.2, method='nearest')

# Works:
tmp    = ds1.sel(time=slice('2016-05-02 10:00', '2016-05-02 12:00'))
dsloc2 = tmp.sel(longitude=4.9, latitude=51.2, method='nearest')

结果如下：

^{pr2}$