html_url,issue_url,id,node_id,user,created_at,updated_at,author_association,body,reactions,performed_via_github_app,issue https://github.com/pydata/xarray/issues/979#issuecomment-457281032,https://api.github.com/repos/pydata/xarray/issues/979,457281032,MDEyOklzc3VlQ29tbWVudDQ1NzI4MTAzMg==,1217238,2019-01-24T17:19:30Z,2019-01-24T17:19:30Z,MEMBER,"I think dask.array handles this differing chunk sizes better these days, so perhaps this is no longer necessary.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,172291585 https://github.com/pydata/xarray/issues/979#issuecomment-264510264,https://api.github.com/repos/pydata/xarray/issues/979,264510264,MDEyOklzc3VlQ29tbWVudDI2NDUxMDI2NA==,5356122,2016-12-02T17:23:46Z,2016-12-02T17:23:46Z,MEMBER,"As an end user, it would be really nice to not have to worry about chunks at all. I'd like to write the same code in xarray using Numpy and have it do the right thing in dask transparently. It seems like dask is moving in this direction (see Automatic blocksize for read_csv dask/dask#1147). Agree with @shoyer that these features belong in dask.","{""total_count"": 1, ""+1"": 1, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,172291585 https://github.com/pydata/xarray/issues/979#issuecomment-241232491,https://api.github.com/repos/pydata/xarray/issues/979,241232491,MDEyOklzc3VlQ29tbWVudDI0MTIzMjQ5MQ==,1217238,2016-08-21T00:55:11Z,2016-08-21T00:55:11Z,MEMBER,"I agree that it would make sense for `xarray.align` to unify chunks in dask arrays, but the documentation is actually a little out of date here: dask.array does now do some minimal automatic rechunking (see [`unify_chunks`](https://github.com/dask/dask/blob/112b541cbb81db04a5753d80da0217aa7f31feeb/dask/array/core.py#L1629) for details). Also, dask array functions, at least those that use [`elemwise`](https://github.com/dask/dask/blob/112b541cbb81db04a5753d80da0217aa7f31feeb/dask/array/core.py#L2195), do automatically coerce NumPy arrays into dask arrays. So adding a tiny numpy array to a huge dask array _does_ currently do the right thing. As you can see, the automatic rechunking algorithm that dask.array currently uses is super simple: it only reconciles chunks when one array is unchunked. I'm certainly open to more sophisticated options for automatic rechunking (see https://github.com/dask/dask/issues/111), but either way I'd prefer to keep as much of this logic on the dask side as possible. Ideally, we'd simply call `dask.array.unify_chunks` passing in the named dimensions for each array. ","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,172291585