I have a string column foo in my DataFrame. I need to create a new column bar, whose values are derived from corresponding foo values by a sequence of string-processing operations - a bunch of str.split()s and str.join()s in this particular case.
What is the most efficient way to do this?
Take a look at the vectorized string methods of pandas dataframes. http://pandas.pydata.org/pandas-docs/dev/text.html#text-string-methods
# You can call whatever vectorized string methods on the RHS
df['bar'] = df['foo']
eg.
df = pd.DataFrame(['a c', 'b d'], columns=['foo'])
df['bar'] = df['foo'].str.split(' ').str.join('-')
print(df)
yields
foo bar
0 a c a-c
1 b d b-d
Pandas can do this for you. A simple example might look like:
foo = ["this", "is an", "example!"]
df = pd.DataFrame({'foo':foo})
df['upper_bar'] = df.foo.str.upper()
df['lower_bar'] = df.foo.str.lower()
df['split_bar'] = df.foo.str.split('_')
print(df)
which will give you
foo upper_bar lower_bar split_bar
0 this THIS this [this]
1 is an IS AN is an [is an]
2 example! EXAMPLE! example! [example!]
See the link above from Alex
