How to select data in xarray Dataset with multiple slices once？

[forestbat]

I have a xarray dataset which looks like this:

Dimensions:    (time: 24, longitude: 701, latitude: 701)
Coordinates:
  * time       (time) datetime64[ns] 192B 2023-06-01 ... 2023-06-01T23:00:00
  * longitude  (longitude) float32 3kB 70.0 70.1 70.2 70.3 ... 139.8 139.9 140.0
  * latitude   (latitude) float32 3kB 65.0 64.9 64.8 64.7 ... -4.8 -4.9 -5.0

And I have some lists of longitudes and latitudes such as:

bboxes = [[122.3, 122.9, 40.3, 39.8], [-124.1, -123.7, 42.4, 42.1]]

If there is only a list, I can select data in dataset with this:

bbox = [122.3, 122.9, 40.3, 39.8]
res = dataset.sel(longitude=slice(bbox[0], bbox[1]), latitude=slice(bbox[2], bbox[3]))

However there are probably hundreds of lists in bboxes, so selecting data from these slices became a difficult task——if I use foreach and xarray.merge to complete it，running speed will be bad.
So how to read data from dataset quickly and elegantly?

[scottyhq]

selecting data from these slices became a difficult task——if I use foreach and xarray.merge to complete it，running speed will be bad.

There is really nothing inherently wrong with a for loop if it accomplishes what you need! The dataset in your example is quite small, but if you really are concerned with scalability here is some discussion that might be of interest https://discourse.pangeo.io/t/advice-for-scalable-raster-vector-extraction/4129/18

So how to read data from dataset quickly and elegantly?

The answer here depends on what you ultimately want to do with your subsets of data. Do you just want summary statistics? The Xvec library provides an elegant approach if you convert your bboxes to a vector index of Polygons https://xvec.readthedocs.io/en/stable/zonal_stats.html.