Drop rows with a 'question mark' value in any column in a pandas dataframe

Question

I want to remove all rows (or take all rows without) a question mark symbol in any column. I also want to change the elements to float type.

Input:

X Y Z
0 1 ?
1 2 3
? ? 4
4 4 4
? 2 5

Output:

X Y Z
1 2 3
4 4 4

Preferably using pandas dataframe operations.

Answer 1

You can try first find string ? in columns, create boolean mask and last filter rows - use boolean indexing. If you need convert columns to float, use astype:

print ~((df['X'] == '?' )  (df['Y'] == '?' ) | (df['Z'] == '?' ))
0    False
1     True
2    False
3     True
4    False
dtype: bool


df1 = df[~((df['X'] == '?' ) | (df['Y'] == '?' ) | (df['Z'] == '?' ))].astype(float)
print df1
   X  Y  Z
1  1  2  3
3  4  4  4

print df1.dtypes
X    float64
Y    float64
Z    float64
dtype: object

Or you can try:

df['X'] = pd.to_numeric(df['X'], errors='coerce')
df['Y'] = pd.to_numeric(df['Y'], errors='coerce')
df['Z'] = pd.to_numeric(df['Z'], errors='coerce')
print df
    X   Y   Z
0   0   1 NaN
1   1   2   3
2 NaN NaN   4
3   4   4   4
4 NaN   2   5
print ((df['X'].notnull() ) & (df['Y'].notnull() ) & (df['Z'].notnull() ))
0    False
1     True
2    False
3     True
4    False
dtype: bool

print df[ ((df['X'].notnull() ) & (df['Y'].notnull() ) & (df['Z'].notnull() )) ].astype(float)
   X  Y  Z
1  1  2  3
3  4  4  4

Better is use:

df = df[(df != '?').all(axis=1)]

Or:

df = df[~(df == '?').any(axis=1)]

Answer 2

You can try replacing ? with null values

import numpy as np

data = df.replace("?", "np.Nan")

if you want to replace particular column try this:

data = df["column name"].replace("?", "np.Nan")