至少有 1 个 Null 值的列名列表以及与每列 pandas 对应的 Null 值总数

Question

Hi I have a code which prints column names along with null values in columns:

A    B     C    D
1    1     4    NAN
2    2     5     NAN
3    NAN   6     NAN

My Code

[IN]res = list(df.isnull().sum().items())
[IN]print(res)

Current Output

[('A', 0), ('B', 1), ('C', 0), ('D', 3)]

Expected output:

[('B', 1), ('D', 3)]

So basically I wish to remove columns where there are 0 null values and return only columns with at least 1 null value.

Answer 1

First idea is use boolean indexing :

s = df.isnull().sum()
res = list(s[s > 0].items())
print (res)
[('B', 1), ('D', 3)]

Or filter using callable :

res = list(df.isnull().sum()[lambda x: x > 0].items())

Or filter in list comprehension:

res = [(k, v) for k, v in df.isnull().sum().items() if v > 0]