如何总结列中包含特定字符串的行中的所有值？

Question

So I have a dataframe of categories of venues in each neighbourhood. It looks like:

The values in each row represent the no. of each venue in the specific neighbourhood.

I want to find out the total number of restaurants in each neighbourhood. To do so, I know I have to sum up the values in a row where the column contains the string "Restaurant".

I've tried using str.contains function but that sums up True cases - how many times a column containing the string restaurant has a value >0 in that row. But instead, what I'd like is, to sum up, the total no. of restaurants in the neighbourhood instead.

Answer 1

Here's a way to do that:

df = pd.DataFrame({"restaurant_a": [1,2,3], "shop": [2,3,4], "restaurant_b": [4,5,6]})
df["sum_rest"] = df[[x for x in df.columns if "restaurant" in x]].sum(axis = "columns")
df

The result is:

   restaurant_a  shop  restaurant_b  sum_rest
0             1     2             4         5
1             2     3             5         7
2             3     4             6         9

Answer 2

You can use pd.Index.str.contains with df.loc here.

df['sum_rest'] = df.loc[:,df.columns.str.contains('Restaurant')].sum(axis=1)

Answer 3

Define a list of columns containing "Restaurant":

lr = ["Afgan Restaurant", "American Restaurant", "Argentinian Restaurant"]

Then parse the result and put it in a column:

df["sum_restaurant"] = df.loc[:, columns=lr].apply(lambda row : np.sum(row.to_numpy()))