I need to find out the datatype of columns with null values only.
df.columns[df_test.isnull().sum()>0]
is giving me the list of columns which has null values , but i need the list of columns with datatypes.
Can you please help !!!
Add a comment
|
2 Answers
Based on what I understand, you can use loc directly without using any loop.
df_test.loc[:,df_test.isnull().sum()>0].dtypes
An example below:
df = pd.util.testing.makeMissingDataframe().fillna({"A":0}).head(5)
print(df)
A B C D
UxNqePUQXH 0.449476 0.713696 1.255964 -0.458908
qBsuuVjimE 0.840831 0.439097 1.073890 1.252807
ptZvWjEdUV 0.377418 0.061335 0.960726 1.140141
LO54c7vqq1 -1.708196 -0.483487 0.831404 NaN
Pc7I7gTRZY NaN -1.400275 0.045297 -0.322303
>>df.loc[:,df.isnull().sum()>0].dtypes
C float64
D float64
dtype: object
Note that you can also explore any:
df_test.loc[:,df_test.isnull().any()].dtypes.
1 Comment
Aniruddha Pal
Thank you so much. Really appreciate it.
import pandas as pd
for x in df:
if df[x].isna().any():
print(df[x].dtype)
that is if the column contains a single null value. if you only want the datatype if ALL the values in the column are null, you can do the below
import pandas as pd
for x in df:
if df[x].isna().all():
print(df[x].dtype)
