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
143107776,MDU6SXNzdWUxNDMxMDc3NzY=,801,dictionaries in labeled indexing with letter values,11950875,closed,0,,,2,2016-03-23T23:44:02Z,2016-03-24T16:05:23Z,2016-03-24T16:05:11Z,CONTRIBUTOR,,,,"An exception is raised when a letter is used in a dictionary as a labelled index. I believe this use of letter values is acceptable according to #291.

The error is:

```
ValueError: invalid literal for int() with base 10: 'b'
```

The work around is simple at least, the .loc keyword works as expected with dictionary labelled indexing.

```
test = np.random.randn(3, 3)
test_xr = xr.DataArray(test, coords=[[1, 2, 3], ['a', 'b', 'c']], dims=['inter', 'alpha'])
test_xr.loc[dict(inter=1)]
test_xr[dict(inter=1)]
test_xr.loc[dict(alpha='b')]
# Throws a ValueError exception
test_xr[dict(alpha='b')]
```
","{""url"": ""https://api.github.com/repos/pydata/xarray/issues/801/reactions"", ""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,completed,13221727,issue
132535802,MDU6SXNzdWUxMzI1MzU4MDI=,753,creating DataArray from list of complex DataArray,11950875,closed,0,,,3,2016-02-09T21:13:50Z,2016-02-11T16:11:52Z,2016-02-11T16:11:52Z,CONTRIBUTOR,,,,"The DataArray constructor fails with a TypeError in the following case:

```
c1 = xr.DataArray(np.array((1, 2), dtype=np.complex))
c2 = xr.DataArray(np.array((3, 4), dtype=np.complex))
test = xr.DataArray([c1, c2])
```

The constructor works fine if c1 and c2 are np.arrays:

```
c1 = np.array((1, 2), dtype=np.complex)
c2 = np.array((3, 4), dtype=np.complex)
test = xr.DataArray([c1,c2])
```

and constructing the array with concatenation also works:

```
c1 = xr.DataArray(np.array((1, 2), dtype=np.complex))
c2 = xr.DataArray(np.array((3, 4), dtype=np.complex))
test = xr.concat([c1, c2], 'new')
```

Any thoughts on how to fix this?
","{""url"": ""https://api.github.com/repos/pydata/xarray/issues/753/reactions"", ""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,completed,13221727,issue
130767608,MDExOlB1bGxSZXF1ZXN0NTgwMjY4NjY=,743,BUG: reindexing an array with complex values should keep dtype=complex,11950875,closed,0,,,6,2016-02-02T18:11:48Z,2016-02-04T23:13:09Z,2016-02-04T23:12:37Z,CONTRIBUTOR,,0,pydata/xarray/pulls/743,"Fixes #738 

I have added a simple behaviour for complex numbers in _maybe_promote. This fix means that  _maybe_promote does indeed return (dtype, np.nan + 0j) for complex dtypes, and the complex -> object conversion shown in #738 does not occur.
","{""url"": ""https://api.github.com/repos/pydata/xarray/issues/743/reactions"", ""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,,13221727,pull
130504978,MDU6SXNzdWUxMzA1MDQ5Nzg=,738,reindex_like with tolerance changes the type of returned DataArray ,11950875,closed,0,,,1,2016-02-01T22:30:36Z,2016-02-04T23:12:37Z,2016-02-04T23:12:37Z,CONTRIBUTOR,,,,"I am really excited about all of the new changes to xarray! reindexing with tolerance has great potential to clean up some sort of nasty stuff that I use a lot but have tucked away into a corner of my code. I am not sure I am 100% comfortable with it at the moment, however, because it seems to be changing the types of my data in unexpected ways.

Bei:

```
test = xr.DataArray([1,2,3], coords=[[0.1, 0.2, 0.3]])
test_i = xr.DataArray([5,6,7,8], coords=[[0.21, 0.31, 0.41, 0.51]])

test.dtype
test.reindex_like(test_i, method='pad', tolerance=0.2).dtype
```

the reindexed output has changed from int64 to float64!

Even more concerning is the complex example:

```
test = xr.DataArray(np.array([1,2,3], dtype=np.complex), coords=[[0.1, 0.2, 0.3]])

test.dtype
test.reindex_like(test_i, method='pad', tolerance=0.2).dtype
```

Conversion from complex to object? Oh no. 

Hope this isn't anything too big, let me know if I can help.
Thank
","{""url"": ""https://api.github.com/repos/pydata/xarray/issues/738/reactions"", ""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,completed,13221727,issue