drop rows with equal mult indexes

Question

Is there a short-cut to drop rows with equal multi indexes? In other words, is there any way to compare indexes without convert them to columns? Here goes how I'm doing so far, but I fill there is something better. Thanks in advance.

>>> df
                0         1         2         3
bar one -1.557968 -1.541893  0.804695  0.838684
    bar -0.148502  0.778159  1.349751 -3.127047
baz one -0.959053 -1.162494 -0.235791  1.239830
    two -1.515304 -0.177502 -1.476930 -1.340431
foo foo -0.797550  0.788936  0.126020  0.229524
    two  0.902523  0.717078 -0.038424  0.339487
>>> df = df.reset_index()
>>> cond = (df2['level_0'] == df2['level_1']) 
>>> df = df.drop(df[cond].index.values)
>>> df
  level_0 level_1         0         1         2         3
0     bar     one -1.557968 -1.541893  0.804695  0.838684
2     baz     one -0.959053 -1.162494 -0.235791  1.239830
3     baz     two -1.515304 -0.177502 -1.476930 -1.340431
5     foo     two  0.902523  0.717078 -0.038424  0.339487

Answer 1

You can use df.index.get_level_values() to get multi-level index values directly.

import pandas as pd
import numpy as np

# replicate your data
a = 'bar bar baz baz foo foo'.split()
b = 'one bar one two foo two'.split()
multi_index = pd.MultiIndex.from_tuples(list(zip(a, b)))
df = pd.DataFrame(np.random.randn(6, 5), index=multi_index)

# do the selection
mask = df.index.get_level_values(0) == df.index.get_level_values(1)
df = df.loc[~mask]

              0       1       2       3       4
bar one -0.0646  0.2245 -0.5863 -0.6400  1.4364
baz one  0.6803  1.6834 -1.0671 -1.0762 -0.8407
    two -0.4484  1.3863 -3.0398 -0.0031 -0.9646
foo two  0.0264  1.4345  0.5046  1.8788 -1.2081