I have the following DataFrame:
df = pd.DataFrame({'Label': list('AABCCC'), 'Values': [1,2,3,4,np.nan,8] })
I want to drop those groups that do not have a minimum number of items (one or less) so I tried the following:
f = lambda x: x.Values.count() > 1
df.groupby('Label').filter(f)
However, this raised an error:
Error : 'numpy.ndarray' object has no attribute 'count'
Where did it go wrong?
It seems you have no Values but values column, so need add [] because collision with values function.
Sample:
df = pd.DataFrame ({'values': [1,2,3,4,np.nan,8] })
print (df)
values
0 1.0
1 2.0
2 3.0
3 4.0
4 NaN
5 8.0
#return numpy array
print (df.values)
[[ 1.]
[ 2.]
[ 3.]
[ 4.]
[ nan]
[ 8.]]
#select column values
print (df['values'])
0 1.0
1 2.0
2 3.0
3 4.0
4 NaN
5 8.0
Name: values, dtype: float64
Your solution for me works nice, I also change .Values to ['Values'].
df1 = df.groupby('Label').filter(lambda x: x['Values'].count() > 1)
print (df1)
Label Values
0 A 1.0
1 A 2.0
3 C 4.0
4 C NaN
5 C 8.0
Alternative solution with transform and boolean indexing:
print (df.groupby('Label')['Values'].transform('count'))
0 2.0
1 2.0
2 1.0
3 2.0
4 2.0
5 2.0
Name: Values, dtype: float64
print (df.groupby('Label')['Values'].transform('count') > 1)
0 True
1 True
2 False
3 True
4 True
5 True
Name: Values, dtype: bool
print (df[df.groupby('Label')['Values'].transform('count') > 1])
Label Values
0 A 1.0
1 A 2.0
3 C 4.0
4 C NaN
5 C 8.0
Also check What is the difference between size and count in pandas?
