How to drop an empty level in a multiindex dataframe

Question

I have a df that looks like this:

mux = pd.MultiIndex(levels=[['A', 'B', 'C', 'D']],labels=[[2, 3, 0, 1]])
df = pd.DataFrame({'a':range(4)}, index=mux)

How can I drop the empty level to form a single-index df? I have tried

pd.MultiIndex.droplevel(df, level=0)

But got the following error:

AttributeError: 'DataFrame' object has no attribute 'levels'

Where did I get wrong?

I have also tried the following, but the index matching for some reason gets messed up (eg. the original row for 'A' is matched with 'B'), even it I use sort_index.

df.index=df.index.levels[0]

sort_index(inplace=True) worked.

xiaoshir
– xiaoshir

2018-07-26 12:35:29 +00:00
Commented Jul 26, 2018 at 12:35 — xiaoshir
– xiaoshir, Commented Jul 26, 2018 at 12:35

jezrael · Accepted Answer · 2018-07-26 12:34:42Z

1

Use get_level_values with assign back or with set_index because it is special type of MultiIndex with only one level:

mux = pd.MultiIndex(levels=[['A', 'B', 'C', 'D']],
       labels=[[2, 3, 0, 1]])

df = pd.DataFrame({'a':range(4)}, index=mux)
print (df)
   a
C  0
D  1
A  2
B  3

df.index = df.index.get_level_values(0)
print (df.index )
Index(['C', 'D', 'A', 'B'], dtype='object')

print (df)
   a
C  0
D  1
A  2
B  3

Or:

df = df.set_index(df.index.get_level_values(0))
print (df.index)
Index(['C', 'D', 'A', 'B'], dtype='object')

edited Jul 26, 2018 at 12:34

answered Jul 26, 2018 at 10:55

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

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

4 Comments

xiaoshir Over a year ago

I tried and got the error: ValueError: Must pass non-zero number of levels/labels

jezrael Over a year ago

@edge27 - So problem with real data? In sample for me it working nice. I add output of df before and after. Or need something different?

xiaoshir Over a year ago

Sorry I thought level[0] is equal to get_level_values(0), but apparently the latter takes care of the matching. So yes, it worked. Or if I use level[0], I need to sort_index first. The problem was I forgot inplace=True.

jezrael Over a year ago

@edge27 - No problem, nice day!

Collectives™ on Stack Overflow

How to drop an empty level in a multiindex dataframe

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related