[Python] Sending large graph of size

client=Client()
ds=xr.open_dataset(file_paths[0], chunks='auto')
time_data = ds['ObservTimes'].compute()


As long as the compute operation is performed, the following warning will appear：

UserWarning: Sending large graph of size 23.71 MiB. This may cause some slowdown. Consider loading the data with Dask directly or using futures or delayed objects to embed the data into the graph without repetition


In theory, the compute operation is put into memory, so why is it embedded in the graph?

The value of ds['ObservTimes'] is as follows：

array(['201301010001', '201301010002', '201301010003', ...,
'201312312358', '201312312359', '201401010000'], dtype='<U12')




Why does this happen? It seems that the values are embedded in the dask graph, but this does not happen in the xarray example on the official documentation.

The following is the structure of the original data

Continue reading...