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
212561278,MDU6SXNzdWUyMTI1NjEyNzg=,1301,open_mfdataset() significantly slower on 0.9.1 vs. 0.8.2,10554254,closed,0,,,17,2017-03-07T21:16:53Z,2017-11-16T15:02:48Z,2017-11-16T15:02:00Z,NONE,,,,"I noticed a big speed discrepancy between xarray versions 0.8.2 and 0.9.1 when using open_mfdataset() on a dataset ~ 1.2 GB in size, consisting of 3 files and using netcdf4 as the engine. 0.8.2 was run first, so this is probably not a disk caching issue.

Test
```
import xarray as xr
import time

start_time = time.time()
ds0 = xr.open_mfdataset('./*.nc')
print(""--- %s seconds ---"" % (time.time() - start_time))
```

Result

xarray==0.8.2, dask==0.11.1, netcdf4==1.2.4
```
--- 0.736030101776 seconds ---
```
xarray==0.9.1, dask==0.13.0, netcdf4==1.2.4

```
--- 52.2800869942 seconds ---
```","{""url"": ""https://api.github.com/repos/pydata/xarray/issues/1301/reactions"", ""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,completed,13221727,issue