Starting with this dataframe df
:
0 1 2
02 en it None
03 en None None
01 nl en fil
There are some missing values. I'm trying to apply a replace function row-wise, eg in pseudocode:
def replace(x):
if 'fil' and 'nl' in row:
x = ''
I know that I can do someting like:
df.apply(f, axis=1)
with a function f
defined like:
def f(x):
if x[0] == 'nl' and x[2] == 'fil':
x[0] = ''
return x
obtaining:
0 1 2
02 en it None
03 en None None
01 en fil
but a priori I don't know the actual positions of the strings through the columns, so I have to search with something like the isin
method, but row-wise.
EDIT: every string can appear anywhere throughout the columns.
You can do something like this:
In [111]:
def func(x):
return x.isin(['fil']).any() & x.isin(['nl']).any()
df.loc[df.apply(func, axis=1)] = df.replace('nl','')
df
Out[111]:
0 1 2
2 en it None
3 en None None
1 en fil
So the function will return True
if both values are present in the row in any position:
In [107]:
df.apply(func, axis=1)
Out[107]:
2 False
3 False
1 True
dtype: bool
You could create boolean indexing based on string comparisons like this
df['0'].str.contains('nl') & df['2'].str.contains('fil')
or since you updated that the columns could change:
df.isin(['fil']).any(axis=1) & df.isin(['nl']).any(axis=1)
Here is the test case:
import pandas as pd
from cStringIO import StringIO
text_file = '''
0 1 2
02 en it None
03 en None None
01 nl en fil
'''
# Read in tabular data
df = pd.read_table(StringIO(text_file), sep='\s+')
print 'Original Data:'
print df
print
# Create boolean index based on text comparison
boolIndx = df.isin(['nl']).any(axis=1) & df.isin(['fil']).any(axis=1)
print 'Example Boolean index:'
print boolIndx
print
# Replace string based on boolean assignment
df.loc[boolIndx] = df.loc[boolIndx].replace('nl', '')
print 'Filtered Data:'
print df
print
Original Data:
0 1 2
2 en it None
3 en None None
1 nl en fil
Example Boolean index:
2 False
3 False
1 True
dtype: bool
Filtered Data:
0 1 2
2 en it None
3 en None None
1 en fil
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.