Converting pandas index to numpy array. Python

Question

I want to convert the index column of a dataframe into an array, but I'm not sure how to go about it. I have already converted the second column to array, but for some reason it doesn't work on the first column as well. Here is how I'm trying to convert the index column:

time = df1.as_matrix(columns = df1.columns[:,0])

But I get

too many indices for array

Here is my dataframe

df1
Out[13]: 
                                 0
2015-11-19 23:59:54.500 -20.186533
2015-11-19 23:59:54.625 -20.272575
2015-11-19 23:59:54.750 -20.185249
2015-11-19 23:59:54.875 -20.247126
2015-11-19 23:59:55.000 -20.205975
2015-11-19 23:59:55.125 -20.281376
2015-11-19 23:59:55.250 -20.238962
2015-11-19 23:59:55.375 -20.300100
2015-11-19 23:59:55.500 -20.311625
2015-11-19 23:59:55.625 -20.264126
2015-11-19 23:59:55.750 -20.266762
2015-11-19 23:59:55.875 -20.224825
2015-11-19 23:59:56.000 -20.211288
2015-11-19 23:59:56.125 -20.163288
2015-11-19 23:59:56.250 -20.254587
2015-11-19 23:59:56.375 -20.125738
2015-11-19 23:59:56.500 -20.146749
2015-11-19 23:59:56.625 -20.161976
2015-11-19 23:59:56.750 -20.126276
2015-11-19 23:59:56.875 -20.082863
2015-11-19 23:59:57.000 -20.030237
2015-11-19 23:59:57.125 -20.098312
2015-11-19 23:59:57.250 -20.146214
2015-11-19 23:59:57.375 -20.030476
2015-11-19 23:59:57.500 -20.018661
2015-11-19 23:59:57.625 -20.029900
2015-11-19 23:59:57.750 -19.970963
2015-11-19 23:59:57.875 -19.994637
2015-11-19 23:59:58.000 -20.097612
2015-11-19 23:59:58.125 -19.952700

For the given example `dataframe`, can you try `time = df.as_matrix(columns = df.columns[:])`? — niraj, Sep 19 '17 at 01:42
Just a note on [`as_matrix`](https://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.as_matrix.html) for future references: "This method is provided for backwards compatibility. Generally, it is recommended to use `.values`." — Brad Solomon, Sep 19 '17 at 01:44
@alexanderson - as of pandas' 0.24.0 - you can access the backing array of a pandas Index with `.array` and `.to_numpy` - please find an updated answer bellow. [pandas 0.24.x release notes](https://pandas.pydata.org/pandas-docs/version/0.24/whatsnew/v0.24.0.html#accessing-the-values-in-a-series-or-index) — mork, Jan 25 '19 at 19:22

score 8 · Accepted Answer · answered Sep 19 '17 at 01:43

8

You can just do df.index.values:

df = pd.DataFrame(index=['a', 'b', 'c'])

df.index.values
# array(['a', 'b', 'c'], dtype=object)

answered Sep 19 '17 at 01:43

Psidom

209,562
33
339
356

Just tried using this method, and I got a numpy.datetime64 which is good, becasue I know how to convert it to datetime from there. Thanks a lot for the help. – alex anderson Sep 19 '17 at 01:50
Cool. Glad it helps. – Psidom Sep 19 '17 at 01:51

score 0 · Answer 2 · answered Sep 19 '17 at 01:47

0

Try time = df1.as_matrix(columns=df1.columns[0:1]). It looks like columns should be a 1-dimensional array (well, actually, an Index), and giving two indices to a 1-dimensional array would give that error.

answered Sep 19 '17 at 01:47

Jeremy McGibbon

3,527
14
22

That gives the second row only, unfortunately. – alex anderson Sep 19 '17 at 01:49
It looks like Psidom's answer should work. – Jeremy McGibbon Sep 19 '17 at 01:50
It did. Thanks for you help though. – alex anderson Sep 19 '17 at 01:51

score 0 · Answer 3 · answered Jan 25 '19 at 19:21

according to pandas 0.24.x release notes: "Series.array and Index.array have been added for extracting the array backing a Series or Index... We haven’t removed or deprecated Series.values or DataFrame.values, but we highly recommend and using .array or .to_numpy()"

Looks like this major version update does exactly what you need :)

Converting pandas index to numpy array. Python

3 Answers3