Python Transform an array

Question

I have this kind of dataframe :

    Variable   Date     Value
0  Variable1  Date1  Valeur 1
1  Variable1  Date2  Valeur 2
2  Variable1  Date3  Valeur 3
3  Variable2  Date4  Valeur 4
4  Variable2  Date5  Valeur 5

I would like to transform it like this :

    Date Variable1 Variable2
0  Date1  Valeur 1      None
1  Date2  Valeur 2      None
2  Date3  Valeur 3      None
3  Date4      None  Valeur 4
4  Date5      None  Valeur 5

How can I do this kind of tranformation in Python with panda or numpy? Thanks for your help

If values < 3 it's going inside Variable 1 else it's going into variable 2 — vincent vorreiter
– vincent vorreiter, Commented Apr 26, 2016 at 14:50

jezrael · Accepted Answer · 2016-04-26 15:04:55Z

4

I think you need pivot with rename_axis (new in pandas 0.18.0) and reset_index:

print df.pivot(index='Date', columns='Variable', values='Value')
        .rename_axis(None, axis=1)
        .reset_index()

    Date Variable1 Variable2
0  Date1  Valeur 1      None
1  Date2  Valeur 2      None
2  Date3  Valeur 3      None
3  Date4      None  Valeur 4
4  Date5      None  Valeur 5

Sample:

import pandas as pd

df = pd.DataFrame({'Variable': {0: 'a', 1: 'a', 2: 'a', 3: 'b', 4: 'b'}, 
                    'Date': {0: pd.Timestamp('2016-02-05 00:00:00'), 
                             1: pd.Timestamp('2016-02-06 00:00:00'),
                             2: pd.Timestamp('2016-02-07 00:00:00'), 
                             3: pd.Timestamp('2016-02-08 00:00:00'), 
                             4: pd.Timestamp('2016-02-09 00:00:00')}, 
                    'Value': {0: 1, 1: 2, 2: 3, 3: 4, 4: 5}},
                    columns=['Variable','Date','Value'])

print df
  Variable       Date  Value
0        a 2016-02-05      1
1        a 2016-02-06      2
2        a 2016-02-07      3
3        b 2016-02-08      4
4        b 2016-02-09      5

print df.pivot(index='Date', columns='Variable', values='Value')
        .rename_axis(None, axis=1)
        .reset_index()

        Date    a    b
0 2016-02-05  1.0  NaN
1 2016-02-06  2.0  NaN
2 2016-02-07  3.0  NaN
3 2016-02-08  NaN  4.0
4 2016-02-09  NaN  5.0

edited Apr 26, 2016 at 15:04

answered Apr 26, 2016 at 14:47

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

vincent vorreiter Over a year ago

Thanks a lot jezrael. I think it's what I needed!

jezrael Over a year ago

Super, first I think you need some special conversion rule ;)

vincent vorreiter Over a year ago

Alright like what? can you give me an exemple please?

vincent vorreiter Over a year ago

Thanks al lot jezrael it's exactly what i needed!

B. M. · Accepted Answer · 2016-04-26 19:29:38Z

0

As a complement, a way to split a column on condition :

df=pd.DataFrame({'Variable':arange(5)},index=pd.date_range('2016/4/26',periods=5))
"""
            Variable
2016-04-26         0
2016-04-27         1
2016-04-28         2
2016-04-29         3
2016-04-30         4
"""

cond=df<3
df[cond].join(df[~cond],lsuffix=1,rsuffix=2)    
"""
            Variable1  Variable2
2016-04-26        0.0        NaN
2016-04-27        1.0        NaN
2016-04-28        2.0        NaN
2016-04-29        NaN        3.0
2016-04-30        NaN        4.0
"""

answered Apr 26, 2016 at 19:29

B. M.

18.7k2 gold badges40 silver badges56 bronze badges

Collectives™ on Stack Overflow

Python Transform an array

2 Answers 2

4 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

4 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related