CODEWITHSUNDEEP
pythondataframepandas
Mads Jensen
Mads Jensen

Reputation: 673

extract data from a MultiIndex'ed DataFrame

I new to pandas and I'm trying to extract data from a MultiIndex'ed DataFrame

is it possible in pandas to select a range of values from a MultiIndex object, e.g. the example DataFrame below I would like to select all of the values from the first level (bar, baz, foo & qux) and the 'one' and 'two' from the second level for all the columns. is that possible?

arrays = [np.array(['bar', 'bar', 'bar', 'baz', 'baz', 'baz', 'foo', 'foo', 'foo', 'qux', 'qux','qux']),
np.array(['one', 'two','three','one', 'two', 'three','one', 'two','three', 'one', 'two','three'])]

df = pd.DataFrame(randn(12, 6), index=arrays)
                  0         1         2         3         4         5
bar one   -0.031447  0.084358 -0.045284 -0.073702 -0.566905 -0.541734
    two   -0.381897  0.422047 -0.527828  0.419984 -0.920186  0.643190
    three  0.082314  2.584901  1.149755 -0.741753  0.696301 -0.132365
baz one    0.637182 -0.210955 -0.329989  0.021509 -0.483201 -1.194727
    two    3.602497 -0.010458  1.734119 -0.332384  0.135082  0.194316
    three -0.293277 -0.144820  0.155034 -0.490092 -0.800939 -0.286902
foo one    1.244119  0.024739  0.500957  0.774194 -3.344261  1.098748
    two   -2.328298 -0.473493  0.881086  0.548408  0.882422 -0.151467
    three  0.512852  1.687702  0.154186 -0.079843  0.116599 -1.330802

Upvotes: 1

Views: 248

Answers (2)

HYRY
HYRY

Reputation: 97261

df.reindex(index=["one", "two"], level=1)

output:

                0         1         2         3         4         5
bar one  0.494206  1.411835  0.047737 -1.750270  0.649225 -0.226546
    two -0.413393  1.686736  0.110594  1.231486  0.135066  2.025476
baz one -1.146431 -0.584855 -1.718917 -0.288630  0.070884 -0.674778
    two  0.957835  1.463544 -0.374227  0.364186  0.259866 -0.019867
foo one  0.300630 -2.648215  0.217727 -1.986657  1.354950 -0.290845
    two  0.046996  1.490452  0.173022 -0.666131 -0.155762 -2.229876
qux one  0.177816 -0.097909  1.360481 -0.619087 -0.026084 -0.512052
    two  0.589484  1.190523  0.759126 -0.380245  1.416895  0.373932

Upvotes: 5

Zelazny7
Zelazny7

Reputation: 40618

I don't think there's an easy way to do this just yet. I think version 0.11 might be addressing this. In the meantime, I find it easiest to just swap the index levels and use .ix:

In [30]: df.swaplevel(1,0).ix[['one','two']]
Out[30]:
                0         1         2         3         4         5
one bar -0.388720 -0.265593 -0.625708 -0.372266  0.425832  1.180458
two bar -2.000612 -0.308223  0.236815  0.327834 -0.837164 -0.525281
one baz  0.039961  1.821628 -0.681662 -0.300518  0.495441  1.204195
two baz -0.243946  0.757397 -1.157399  0.044435  0.325952 -1.613384
one foo -0.322431  0.539729 -0.081879  1.129230 -0.324549 -1.252578
two foo  1.127411 -1.957569 -0.284485  0.604807 -0.326727 -1.025945
one qux -0.217390 -3.194743  0.939121 -0.741615 -1.640926 -0.157857
two qux  0.513316  1.568216 -1.702408  0.955211 -0.017799 -0.718302

Upvotes: 0

Related Questions