CODEWITHSUNDEEP
pythonpandasnumpydataframe
user11509999
user11509999

Reputation:

Python Dataframe: Remove digits in odd position and even position from df rows?

I have a df with two columns number ,type. I'm trying to remove the numbers in odd position if type is change and condition is odd.

And Similarly numbers in even position if type is change and condition is even.

For number - (7,14,21,28) without numbers in odd position - 7,21.

Similarly - (3,6,5,8) without numbers in even position - 6,8

df

    df= pd.DataFrame(data = {'number' : ['7,14,21,28', '2,5', '3,6,5,8 ', '4', '3,10,17,24,31'], 'type' : ['change', 'dont_change', 'change', 'dont_change', 'change'], 'condition':['odd','even','even','odd','odd']}) 


    number          type             condition
0   7,14,21,28      change             odd
1   2,5             dont_change        even
2   3,6,5,8         change             even 
3   4               dont_change        odd
4   3,10,17,24,31   change             odd

My excepted output:

    number          type             condition
0   7,21            change             odd
1   2,5             dont_change        even
2   6,8             change             even 
3   4               dont_change        odd
4   3,17,31         change             odd

What i have tried:

I have two different code for even and odd. Without adding type - condition - change/dont_change

string slicing with step = 2


even

df.number = np.where(df['condition'].eq('odd'), df.number.str.split(',').str[1::2].str.join(','), df['number'])


odd

df.number = np.where(df['condition'].eq('odd'), df.number.str.split(',').str[::2].str.join(','), df['number'])

How do i proceed from this?

Upvotes: 0

Views: 214

Answers (2)

Balaji Ambresh
Balaji Ambresh

Reputation: 5012

Could you give this a shot?

def filter_numbers(row):
    if row['type'] == 'change':
        if row['condition'] == 'odd':
            return ','.join(row['number'].split(',')[::2])
        else:
            return ','.join(row['number'].split(',')[1::2])
    return row['number']

df['number'] = df.apply(filter_numbers, axis=1)

print(df)

Output:

    number         type condition
0     7,21       change       odd
1      2,5  dont_change      even
2     6,8        change      even
3        4  dont_change       odd
4  3,17,31       change       odd

Upvotes: 0

BENY
BENY

Reputation: 323316

Let us try np.select 

cond1=df['condition'].eq('odd') & df.type.eq('change')
cond2=df['condition'].eq('even') & df.type.eq('change')
v1=df.number.str.split(',').str[::2].str.join(',')
v2=df.number.str.split(',').str[1::2].str.join(',')
df.number=np.select([cond1,cond2], [v1,v2], default = df.number)
df
Out[417]: 
    number         type condition
0     7,21       change       odd
1      2,5  dont_change      even
2     6,8        change      even
3        4  dont_change       odd
4  3,17,31       change       odd

Upvotes: 1

Related Questions