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/471#issuecomment-122333205,https://api.github.com/repos/pydata/xarray/issues/471,122333205,MDEyOklzc3VlQ29tbWVudDEyMjMzMzIwNQ==,1217238,2015-07-17T16:27:50Z,2015-07-17T16:27:50Z,MEMBER,"Awesome, thanks!

This is where you'll find the decoding logic:
https://github.com/xray/xray/blob/v0.5.2/xray/conventions.py#L663
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,94966000
https://github.com/pydata/xarray/issues/471#issuecomment-122332344,https://api.github.com/repos/pydata/xarray/issues/471,122332344,MDEyOklzc3VlQ29tbWVudDEyMjMzMjM0NA==,1050278,2015-07-17T16:23:34Z,2015-07-17T16:23:34Z,CONTRIBUTOR,"Sounds like a good solution.  I'll work on a PR.
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,94966000
https://github.com/pydata/xarray/issues/471#issuecomment-122331414,https://api.github.com/repos/pydata/xarray/issues/471,122331414,MDEyOklzc3VlQ29tbWVudDEyMjMzMTQxNA==,1217238,2015-07-17T16:18:58Z,2015-07-17T16:18:58Z,MEMBER,"What about issuing a warning, and then decoding the duplicate values to NaN anyways? I'm usually not a fan of warnings, but my guess is that this would be the most helpful behavior for users.
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,94966000
https://github.com/pydata/xarray/issues/471#issuecomment-122294746,https://api.github.com/repos/pydata/xarray/issues/471,122294746,MDEyOklzc3VlQ29tbWVudDEyMjI5NDc0Ng==,1050278,2015-07-17T14:30:15Z,2015-07-17T14:30:15Z,CONTRIBUTOR,"Yes, the two values in the missing_value attribute indicate two classes of data (not collected vs below minimum detectable threshold) and these data can be access by setting mask_and_scale=False but this also results in a valid data being returned without scaling which makes it less useful.  

My question is how should xray should handle these cases?  Either replace all instances of the values in missing_value with NaN or raise a error message stating that multiple missing_values are not supported similar?  I'd be happy to create a PR implementing either case.
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,94966000
https://github.com/pydata/xarray/issues/471#issuecomment-121283168,https://api.github.com/repos/pydata/xarray/issues/471,121283168,MDEyOklzc3VlQ29tbWVudDEyMTI4MzE2OA==,1217238,2015-07-14T15:32:54Z,2015-07-14T15:32:54Z,MEMBER,"You can still read these files if you set mask_and_scale=False in open_dataset -- it just won't decode the missing values for you.

Arguably, we should raise an error message saying just that.

From a practical perspective, are you interested in the different types of these missing values? Presumably there is some reason why they were coded differently in the first place. 
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,94966000