Add sel_points for point-wise indexing by label

[shoyer]

xref #475

Example usage:

In [1]: da = xray.DataArray(np.arange(56).reshape((7, 8)),
   ...:                     coords={'x': list('abcdefg'),
   ...:                             'y': 10 * np.arange(8)},
   ...:                     dims=['x', 'y'])
   ...:

In [2]: da
Out[2]:
<xray.DataArray (x: 7, y: 8)>
array([[ 0,  1,  2,  3,  4,  5,  6,  7],
       [ 8,  9, 10, 11, 12, 13, 14, 15],
       [16, 17, 18, 19, 20, 21, 22, 23],
       [24, 25, 26, 27, 28, 29, 30, 31],
       [32, 33, 34, 35, 36, 37, 38, 39],
       [40, 41, 42, 43, 44, 45, 46, 47],
       [48, 49, 50, 51, 52, 53, 54, 55]])
Coordinates:
* y        (y) int64 0 10 20 30 40 50 60 70
* x        (x) |S1 'a' 'b' 'c' 'd' 'e' 'f' 'g'

# we can index by position along each dimension
In [3]: da.isel_points(x=[0, 1, 6], y=[0, 1, 0], dim='points')
Out[3]:
<xray.DataArray (points: 3)>
array([ 0,  9, 48])
Coordinates:
    y        (points) int64 0 10 0
    x        (points) |S1 'a' 'b' 'g'
  * points   (points) int64 0 1 2

# or equivalently by label
In [4]: da.sel_points(x=['a', 'b', 'g'], y=[0, 10, 0], dim='points')
Out[4]:
<xray.DataArray (points: 3)>
array([ 0,  9, 48])
Coordinates:
    y        (points) int64 0 10 0
    x        (points) |S1 'a' 'b' 'g'
  * points   (points) int64 0 1 2
Bug fixes

cc @jhamman

[jhamman]

I know these bullets are straight from pandas but "default" isn't a valid keyword here. I think this may be clearer as * None (default): only exact matches.

[shoyer]

good point, I'll update.

[jhamman]

This looks pretty good to me. Are you basically letting pandas do all the testing in terms of the method and range of the indexers? I'm wondering if we need to target any corner cases that extend beyond how pandas does the indexer mapping.

[shoyer]

Are you basically letting pandas do all the testing in terms of the method and range of the indexers? I'm wondering if we need to target any corner cases that extend beyond how pandas does the indexer mapping.

This has been my strategy. Pandas has lots of tests for the exact behavior of get_indexer.

[shoyer]

@jhamman any other comments? If not, I'll merge this shortly.

[jhamman]

No, I think this is good to go.

[jhamman]

Are there going to be merge conflicts with #512? Maybe merge this then update that PR accordingly?

[WeatherGod]

you have two test methods with the same name