issue_comments

Table actions

3 rows where issue = 232350436 and user = 1217238 sorted by updated_at descending

✎ View and edit SQL

This data as json, CSV (advanced)

Suggested facets: reactions, created_at (date), updated_at (date)

user 1

  • shoyer · 3 ✖

issue 1

  • xr.concat loses coordinate dtype information with recarrays in 0.9 · 3 ✖

author_association 1

id html_url issue_url node_id user created_at updated_at ▲ author_association body reactions performed_via_github_app issue
388660444 https://github.com/pydata/xarray/issues/1434#issuecomment-388660444 https://api.github.com/repos/pydata/xarray/issues/1434 MDEyOklzc3VlQ29tbWVudDM4ODY2MDQ0NA== shoyer 1217238 2018-05-13T22:27:28Z 2018-05-13T22:27:28Z MEMBER

@ctw I'm not sure why #1438 was closed (I think that was by mistake), but we would still need a fix along those lines to fix this issue.

 
{
    "total_count": 0,
    "+1": 0,
    "-1": 0,
    "laugh": 0,
    "hooray": 0,
    "confused": 0,
    "heart": 0,
    "rocket": 0,
    "eyes": 0
}
   xr.concat loses coordinate dtype information with recarrays in 0.9 232350436
305036355 https://github.com/pydata/xarray/issues/1434#issuecomment-305036355 https://api.github.com/repos/pydata/xarray/issues/1434 MDEyOklzc3VlQ29tbWVudDMwNTAzNjM1NQ== shoyer 1217238 2017-05-30T23:20:14Z 2017-05-30T23:20:14Z MEMBER

Is there an advantage to using pandas for concatenation over numpy.concatenate?

Yes, otherwise we loose pandas (d)type information. For example, a MultiIndex gets turned into a dtype=object array of tuples.

 
{
    "total_count": 1,
    "+1": 1,
    "-1": 0,
    "laugh": 0,
    "hooray": 0,
    "confused": 0,
    "heart": 0,
    "rocket": 0,
    "eyes": 0
}
   xr.concat loses coordinate dtype information with recarrays in 0.9 232350436
305034234 https://github.com/pydata/xarray/issues/1434#issuecomment-305034234 https://api.github.com/repos/pydata/xarray/issues/1434 MDEyOklzc3VlQ29tbWVudDMwNTAzNDIzNA== shoyer 1217238 2017-05-30T23:08:48Z 2017-05-30T23:08:48Z MEMBER

Thanks for the report, I'm marking this as a bug.

The issue is with these lines in IndexVariable.concat. We defer to pandas for concatenating coordinate labels, but pandas doesn't handle structured dtypes properly (it converts them to dtype=object). This is part of why we have PandasIndexAdapter.

It seems like we now need some more complex logic for toggling between pandas Index.append() and numpy.concatenate, based on the Index dtypes.

 
{
    "total_count": 0,
    "+1": 0,
    "-1": 0,
    "laugh": 0,
    "hooray": 0,
    "confused": 0,
    "heart": 0,
    "rocket": 0,
    "eyes": 0
}
   xr.concat loses coordinate dtype information with recarrays in 0.9 232350436

Advanced export

JSON shape: default, array, newline-delimited, object

CSV options: 

CREATE TABLE [issue_comments] (
   [html_url] TEXT,
   [issue_url] TEXT,
   [id] INTEGER PRIMARY KEY,
   [node_id] TEXT,
   [user] INTEGER REFERENCES [users]([id]),
   [created_at] TEXT,
   [updated_at] TEXT,
   [author_association] TEXT,
   [body] TEXT,
   [reactions] TEXT,
   [performed_via_github_app] TEXT,
   [issue] INTEGER REFERENCES [issues]([id])
);
CREATE INDEX [idx_issue_comments_issue]
    ON [issue_comments] ([issue]);
CREATE INDEX [idx_issue_comments_user]
    ON [issue_comments] ([user]);