Mapping a Series with a NumPy array -- dimensionality issue?

Question

When I'm using 2d array maps, everything works fine. When I start using 1d arrray's this error occurs; IndexError: unsupported iterator index. This is the error I'm talking about:

In [426]: y = Series( [0,1,0,1] )
In [427]: arr1 = np.array( [10,20] )
In [428]: arr2 = np.array( [[10,20],[30,40]] )
In [429]: arr2[ y, y ]
Out[429]: array([10, 40, 10, 40])
In [430]: arr1[ y ]

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
<ipython-input-430-25b98edce1f3> in <module>()
----> 1 arr1[ y ]

IndexError: unsupported iterator index

I'm using the latest Anaconda distribution with NumPy 1.8.1. Maybe this is related to a NumPy bug discussed here? Could anybody tell me what is causing this error?

Jeff · Accepted Answer · 2014-07-25 14:46:08Z

2

You need to either convert the Series to a array, or vice-versa. Indexers must be 1-d for a 1-d object.

In [11]: arr1[y.values]
Out[11]: array([10, 20, 10, 20])

In [12]: Series(arr1)[y]
Out[12]: 
0    10
1    20
0    10
1    20
dtype: int64

answered Jul 25, 2014 at 14:46

Jeff

130k21 gold badges223 silver badges189 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

JohnE Over a year ago

Thanks, that's perfect. Looks like I need to work on my understanding of numpy indexing.

Jeff Over a year ago

if need be you can convert a 2-d to a 1-d with arr.ravel()

JohnE Over a year ago

Good to know. I was also working around with arr2[y,0] but wanted to figure out what was happening in case I was not understanding how numpy works (which I obv wasn't!)

Collectives™ on Stack Overflow

Mapping a Series with a NumPy array -- dimensionality issue?

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related