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
659142789,MDU6SXNzdWU2NTkxNDI3ODk=,4236,Allow passing args to preprocess function in open_mfdataset,8098361,closed,0,,,7,2020-07-17T10:52:14Z,2023-09-12T15:59:45Z,2023-09-12T15:59:45Z,NONE,,,,"For a set of netcdf files I'm opening with `open_mfdataset` I'd also like to pass a couple of extra arguments to the `preprocess` function. At the moment the `Dataset` seems to be the only arg that the `preprocess` function accepts.

The netcdf files have dimensions (time, lat, lon). It's the time dimension I'd like to cut down during a parallel load eg. using a start/end dayofyear. Each file covers a different year and I'd like to slice out particular dayofyear range within each. I tried calculating dayofyear inside the `preprocess` function and doing `.sel`, which works perfectly fine, but not with the `parallel=True` option. Without the parallel option loading is much slower.
However, if `parallel=True` a pickling error occurs, possibly because I'm using other functions like dateparse, or timedelta inside the preprocess function to calculate the dayofyear (which itself is derived from a ipywidget).
I don't really understand the pickling error, it says; 
```
D:\Anaconda3\lib\pickle.py in save_global(self, obj, name)
    963                 raise PicklingError(
    964                     ""Can't pickle %r: it's not the same object as %s.%s"" %
--> 965                     (obj, module_name, name))
    966 
    967         if self.proto >= 2:

PicklingError: Can't pickle <built-in function input>: it's not the same object as builtins.input
```
I also tried setting the start/end dayofyear as global outside the `preprocess` function and using them inside the function but changing those globals (integers) after the function is defined doesn't seem to alter the reference to them inside the function. I don't think this solution is very elegant anyway.

Many other packages have functions that use callbacks with the option of passing additional arguments. Has this been considered for open_mfdataset?","{""url"": ""https://api.github.com/repos/pydata/xarray/issues/4236/reactions"", ""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,completed,13221727,issue