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/pull/2073#issuecomment-383364259,https://api.github.com/repos/pydata/xarray/issues/2073,383364259,MDEyOklzc3VlQ29tbWVudDM4MzM2NDI1OQ==,16351486,2018-04-22T08:21:20Z,2018-04-22T08:21:20Z,NONE,"When I am a novice, I want to convert multidimensional data of ```pd.Dataframe``` into ```xr.Dataset```. The 'Multi-dimensional data' section is found from the 'Working with pandas' chapter. Then, I skipping the other to this section. this example showed that Xarray is not good at Multi-dimensional data until i discovered ```xr.Dataset.from_dataframe``` i know the mean of the leading paragraph, but it may be obstacles in this the 'Multi-dimensional data' section for novices.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,316540429 https://github.com/pydata/xarray/pull/2073#issuecomment-383355977,https://api.github.com/repos/pydata/xarray/issues/2073,383355977,MDEyOklzc3VlQ29tbWVudDM4MzM1NTk3Nw==,1217238,2018-04-22T05:10:17Z,2018-04-22T05:10:17Z,MEMBER,"I agree that `from_dataframe` should be preferred -- which is exactly why we mention in first on this doc page. I'm not sure this change makes sense, since the leading paragraph introduces this example with ""Both the DataArray and Dataset constructors...""","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,316540429 https://github.com/pydata/xarray/pull/2073#issuecomment-383350237,https://api.github.com/repos/pydata/xarray/issues/2073,383350237,MDEyOklzc3VlQ29tbWVudDM4MzM1MDIzNw==,16351486,2018-04-22T02:24:17Z,2018-04-22T02:58:49Z,NONE," No, these should not An example --- ```python # creat dataframe index = pd.MultiIndex.from_arrays([['a', 'a', 'b'], [0, 1, 2]], names=['one', 'two']) df = pd.DataFrame({'x': 1, 'y': 2}, index=index) # display df ```
x y
one two
a 0 1 2
1 1 2
b 2 1 2
--- ```python # use xr.Dataset transform xr.Dataset(df) ``` Dimensions: (dim_0: 3) Coordinates: * dim_0 (dim_0) MultiIndex - one (dim_0) object 'a' 'a' 'b' - two (dim_0) int64 0 1 2 Data variables: x (dim_0) int64 1 1 1 y (dim_0) int64 2 2 2 --- ```python # use xr.Dataset.from_dataframe transform xr.Dataset.from_dataframe(df) ``` Dimensions: (one: 2, two: 3) Coordinates: * one (one) object 'a' 'b' * two (two) int64 0 1 2 Data variables: x (one, two) float64 1.0 1.0 nan nan nan 1.0 y (one, two) float64 2.0 2.0 nan nan nan 2.0 --- From the above example,```xr.Dataset.from_dataframe``` is a better representation of the coordinate than ```xr.Dataset```from pandas.dataframe ","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,316540429 https://github.com/pydata/xarray/pull/2073#issuecomment-383348212,https://api.github.com/repos/pydata/xarray/issues/2073,383348212,MDEyOklzc3VlQ29tbWVudDM4MzM0ODIxMg==,5635139,2018-04-22T01:51:08Z,2018-04-22T01:51:08Z,MEMBER,Should these return the same result? ,"{""total_count"": 1, ""+1"": 1, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,316540429