Pandas alternate way to add new column with lot of conditions other than apply

问题

I have two dataframes, let's say df and map_dum. Here is the df.

>>> print(df)
    sales
0       5
1      10
2       9
3       7
4       1
5       1
6      -1
7       2
8       9
9       8
10      1
11      3
12     10
13     -2
14      8
15      5
16      9
17      6
18     10
19     -1
20      5
21      3

And here is for the map_dum.

>>> print(map_dum)
   class  more_than_or_equal_to  less_than
0     -1                  -1000          0
1      1                      0          2
2      2                      2          4
3      3                      4          6
4      4                      6          8
5      5                      8         10
6      6                     10       1000

My goal is to add new column to the df, column class. In order to do so, I have to check the value in df['sales'] lies in between which values in map_dum. For example if I want to know the class for the first row in df['sales'], 5, then the class would be 3. The final output would like below.

>>> print(df)
    sales  class
0       5      3
1      10      6
2       9      5
3       7      4
4       1      1
5       1      1
6      -1     -1
7       2      2
8       9      5
9       8      5
10      1      1
11      3      2
12     10      6
13     -2     -1
14      8      5
15      5      3
16      9      5
17      6      4
18     10      6
19     -1     -1
20      5      3
21      3      2

Currently, I am using apply to solve this, however, it is very slow since my dataset is quite huge.

def add_class(sales, mapping, lower_limit, upper_limit):
    result = mapping.loc[((mapping[lower_limit]<=sales)&(mapping[upper_limit]>sales)), 'class'].tolist()[0]
    return result

df['class'] = df['sales'].apply(lambda sales: add_class(sales, map_dum, 'more_than_or_equal_to', 'less_than'))

Hence, performance does matter in my case. Any other way to add the class column to the df without violating the criteria, something like vectorization solution? Thanks for any help!

回答1:

I think you need cut:

bins = [-1000, 0, 2, 4, 6, 8, 10, 1000]
labels=[-1,1,2,3,4,5,6]
df['class'] = pd.cut(df['sales'], bins=bins, labels=labels, right=False)
print (df)
    sales class
0       5     3
1      10     6
2       9     5
3       7     4
4       1     1
5       1     1
6      -1    -1
7       2     2
8       9     5
9       8     5
10      1     1
11      3     2
12     10     6
13     -2    -1
14      8     5
15      5     3
16      9     5
17      6     4
18     10     6
19     -1    -1
20      5     3
21      3     2

For dynamic add values from map_dum use:

bins = [map_dum['more_than_or_equal_to'].iat[0]] + map_dum['less_than'].tolist()
labels= map_dum['class']
df['class'] = pd.cut(df['sales'], bins=bins, labels=labels, right=False)
print (df)

来源：https://stackoverflow.com/questions/42649224/pandas-alternate-way-to-add-new-column-with-lot-of-conditions-other-than-apply