CODEWITHSUNDEEP
pythonpandasdataframe
Kevin Van Humbeeck
Kevin Van Humbeeck

Reputation: 55

Python Multiple Assignment with Panda Series

I would like to fill 2 destination pandas columns at once for performance reasons. A sample value from the source column is '42844 - News - BENL'. I need to extract '42844' and 'News'. Additional values can be ignored.

df["_site_id"], df["_site_name"], rest = df["website"].apply(lambda x: x.split(" - ", 2))

I tested this code and i get a 'ValueError: too many values to unpack (expected 3)'

Upvotes: 2

Views: 89

Answers (2)

yatu
yatu

Reputation: 88236

For vectorized string operations in pandas you have str. So instead you can use Series.str.split. Also for this assignment to work you have to first "expand" the resulting list into n new columns, so set expand=True:

df["_site_id"], df["_site_name"], _ = df.website.str.split(' - ', 2, expand=True)

Which could also be done by indexing on a list of columns and splatting on the lhs as:

*df[['_site_id', '_site_name']], _ = df.website.str.split(' - ', 2, expand=True)

Note: For multiple assignment to work there must be 3 columns as a result from splitting website, otherwise you'll get a ValueError saying that there are not enough values to unpack.

Upvotes: 3

BENY
BENY

Reputation: 323306

Let us do str.split

df["website"].str.split(" - ", 2, expand=True)
#df=df.join(df["website"].str.split(" - ", 2, expand=True).rename(columns={0:'_site_id', 1:'_site_name'}))

Upvotes: 1

Related Questions