CODEWITHSUNDEEP
pythonpandasnumpydataframedata-structures
Deepak M
Deepak M

Reputation: 1224

Splitting a number and creating new individual columns for each number split using Pandas

Good day,

I have a data frame with a column like this, let us assume with 1000rows but here is a sample:

 A 
12
24
36
48

I wish to split the number into two separate numbers. I want the output to look like this:

 A    B    C
12    1    2
24    2    4
36    3    6
48    4    8

How can I achieve this using Pandas and Numpy? Help would be truly appreciated. Thanks in advance!

Upvotes: 1

Views: 144

Answers (4)

FLab
FLab

Reputation: 7466

Another approach based on splitting every character of the number treated as string:

df = pd.DataFrame([12, 24, 36, 48], columns=['A'])

values = df['A'].values
split = [list(str(el)) for el in values]

out = pd.DataFrame(split, columns=['B', 'C']).astype(int)

which gives:

out
   B  C
0  1  2
1  2  4
2  3  6
3  4  8

Upvotes: 2

jezrael
jezrael

Reputation: 862511

Use floor and mod:

df['B'] = df['A'] // 10
df['C'] = df['A'] % 10

print (df)
    A  B  C
0  12  1  2
1  24  2  4
2  36  3  6
3  48  4  8

If input data are strings is possible indexing by positions by []:

print (df['A'].apply(type))
0    <class 'str'>
1    <class 'str'>
2    <class 'str'>
3    <class 'str'>
Name: A, dtype: object

df['B'] = df['A'].str[0]
df['C'] = df['A'].str[1]
#if necessary convert all columns to integers
df = df.astype(int)
print (df)
    A  B  C
0  12  1  2
1  24  2  4
2  36  3  6
3  48  4  8

Upvotes: 3

EdChum
EdChum

Reputation: 393943

For a df that size use floordiv and mod:

In[141]:
df['B'] = df['A'].floordiv(10)
df['C'] = df['A'].mod(10)
df

Out[141]: 
    A  B  C
0  12  1  2
1  24  2  4
2  36  3  6
3  48  4  8

There are also the numpy equivalents, np.floor_divide and np.mod:

In[142]:
df['B'] = np.floor_divide(df['A'],10)
df['C'] = np.mod(df['A'],10)
df

Out[142]: 
    A  B  C
0  12  1  2
1  24  2  4
2  36  3  6
3  48  4  8

The numpy versions are faster:

%%timeit
df['B'] = df['A'].floordiv(10)
df['C']= df['A'].mod(10)
1000 loops, best of 3: 733 µs per loop

%%timeit
df['B'] = np.floor_divide(df['A'],10)
df['C'] = np.mod(df['A'],10)

1000 loops, best of 3: 491 µs per loop

Upvotes: 3

MaxU - stand with Ukraine
MaxU - stand with Ukraine

Reputation: 210832

In [15]: df.A.astype(str).str.extractall(r'(.)')[0].unstack().astype(np.int8)
Out[15]:
match  0  1
0      1  2
1      2  4
2      3  6
3      4  8

Upvotes: 2

Related Questions