从 Pandas 数据框中获取列名，包括索引名

Question

Assuming that we have a data frame with an index that might have a name:

df = pd.DataFrame({'a':[1,2,3],'b':[3,6,1], 'c':[2,6,0]})
df = df.set_index(['a'])

   b  c
a      
1  3  2
2  6  6

What is the best way to get the column names that will include the index name if it is present.

Calling df.columns.tolist() do not include the index name and return ['b', 'c'] in this case, and I would like to obtain ['a', 'b', 'c'] .

Answer 1

The index can be temporarily reset for the call:

df.reset_index().columns.tolist()

If an empty index name is not to appear in the list, do the reset_index() conditionally:

(df.reset_index() if df.index.name else df).columns.tolist()

Answer 2

For universal solution need filter None if not exist index.name :

df = pd.DataFrame({'a':[1,2,3],'b':[3,6,1], 'c':[2,6,0]})

print ([df.index.name] + df.columns.tolist())
[None, 'a', 'b', 'c']

c = list(filter(None, [df.index.name] + df.columns.tolist()))
print (c)
['a', 'b', 'c']

df = df.set_index(['a'])

c = list(filter(None, [df.index.name] + df.columns.tolist()))
print (c)
['a', 'b', 'c']

Another solution with numpy.insert and difference :

c = np.insert(df.columns, 0, df.index.name).difference([None]).tolist()
print (c)

['a', 'b', 'c']

Answer 3

You can use list with filter after elevating your index via reset_index :

res = list(filter(None, df.reset_index()))

print(res)

['a', 'b', 'c']

Answer 4

I think with more recent versions of pandas this answer might be more concise:

names = list(filter(None, df.index.names + df.columns.values.tolist()))

This works for no index, single column Index, or MultiIndex. It avoids calling reset_index() which has an unnecessary performance hit for such a simple operation.

Answer 5

我想，你是从 group by operation 中得到的，我想如果是这样的话，你需要在最后添加 reset_index() 以通过常规方法获取列名。