Pandas pivot_table dropna 参数未按预期工作

Question

I'm pivoting a dataframe to look at unique value counts within groups. I know some of the grouping columns have null values and I want to include them. I can do this easily with a .groupby([...], dropna=False) but I would like to use .pivot_table as it handles the unstacking, null-filling, totaling, etc all in one function.

Sample Data (taken from python pandas: pivot_table silently drops indices with nans )

a = [['a', 'b', 12, 12, 12], ['a', np.nan, 12.3, 233., 12], ['b', 'a', 123.23, 123, 1], ['a', 'b', 1, 1, 1.]]

df = pd.DataFrame(a, columns=['a', 'b', 'c', 'd', 'e'])

print(df)

   a    b       c      d     e
0  a    b   12.00   12.0  12.0
1  a  NaN   12.30  233.0  12.0
2  b    a  123.23  123.0   1.0
3  a    b    1.00    1.0   1.0

Using .groupby to get desired results

using_groupby = df.groupby([
    "a",
    "b"
], dropna=False).c.nunique().unstack(fill_value=0)

print(using_groupby)



b  a  b  NaN
a           
a  0  2    1
b  1  0    0

Code I expected would yield similar results using .pivot_table

using_pivot_table = df.pivot_table(
    index="a",
    columns="b",
    values="c",
    aggfunc="nunique",
    fill_value=0,
    dropna=False
)

print(using_pivot_table)



b  a  b
a      
a  0  2
b  1  0

Question

Is this a bug in the pivot_table function? Or am I not understanding the use of the dropna param?

Version Info

• Python - 3.8.5
• Pandas - 1.1.3

Answer 1

dropna=False means do not include columns whose entries are all NaN; your issue is that the pivot table is not displaying a column with NaN as the column name. If you change the NaN value to another string, then the pivot table works as expected.

df['b'] = df['b'].fillna('No Value')