id,node_id,number,title,user,state,locked,assignee,milestone,comments,created_at,updated_at,closed_at,author_association,active_lock_reason,draft,pull_request,body,reactions,performed_via_github_app,state_reason,repo,type
645443880,MDU6SXNzdWU2NDU0NDM4ODA=,4180,to_netcdf very slow for some single character data types,7360639,closed,0,,,2,2020-06-25T10:18:53Z,2020-06-26T09:45:39Z,2020-06-26T09:45:39Z,NONE,,,,"Hi,

I'm not entirely sure if this is a bug, my misuse, or a feature request. However, I don't think that this is desired behaviour: saving a dataset with single characters in the initial data type is very slow compared to a less desirable alternative.




Example:
make a smaller fake dataset, save it as it is (tester has dtype '<U1'), then change data type and do it again

``` python
import xarray as xr
import numpy as np
import cProfile


ds = xr.Dataset({'tester':('index',np.full((1000000),'.'))},coords = {'index':np.arange(0,1000000)})


cProfile.run(""""""ds.to_netcdf('somefilename')"""""")


ds.tester.values = ds.tester.values.astype('|S1')
cProfile.run(""""""ds.to_netcdf('somefilename')"""""")
```



I find that the first option takes around 8s and the second around 0.076s. This is a massive difference - my own dataset is much larger than this so I am obliged to save it using the |S1 dtype. However, this is much more difficult to use. I am using this field for quality control flags, and now == '.' doesn't work, so I have to wrap saving and loading these datasets with a function changing the datatype.

So I have a workaround but I don't think it makes sense to be this much slower to save single character strings. I also tried pickling the dataset with '<U1' datatype - that is very fast, but netCDF would be better.




<details><summary>Output of <tt>xr.show_versions()</tt></summary>
python: 3.7.6 | packaged by conda-forge | (default, Jun  1 2020, 18:57:50) 
[GCC 7.5.0]
python-bits: 64
OS: Linux
OS-release: 3.10.0-957.27.2.el7.x86_64
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_GB.UTF-8
LOCALE: en_GB.UTF-8
libhdf5: 1.10.5
libnetcdf: 4.7.4

xarray: 0.15.1
pandas: 1.0.4
numpy: 1.18.5
scipy: 1.4.1
netCDF4: 1.5.3
pydap: None
h5netcdf: None
h5py: 2.10.0
Nio: None
zarr: None
cftime: 1.1.3
nc_time_axis: None
PseudoNetCDF: None
rasterio: None
cfgrib: None
iris: None
bottleneck: 1.3.2
dask: 2.18.1
distributed: 2.18.0
matplotlib: 3.2.1
cartopy: None
seaborn: None
numbagg: None
setuptools: 47.1.1.post20200529
pip: 20.1.1
conda: None
pytest: None
IPython: None
sphinx: None


</details>","{""url"": ""https://api.github.com/repos/pydata/xarray/issues/4180/reactions"", ""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,completed,13221727,issue