Combine Rows in Pandas DataFrame

Question

I have financial performance Indicators for different companies, one row per year. Now I would like to have all the indicators per company over a specific range of years in one row.

Now my data looks similar to this:

import numpy as np
import pandas as pd


startyear = 2014
endyear = 2015

df = pd.DataFrame(np.array([
['AAPL',  2014,  0.2,  0.4,  1.5],
['AAPL',  2015,  0.3,  0.4,  2.0],
['AAPL',  2016,  0.2,  0.3,  1.5],
['GOGL',  2014,  0.4,  0.5,  0.5],
['GOGL',  2015,  0.6,  0.8,  1.0],
['GOGL',  2016,  0.3,  0.5,  2.0]]), 
columns=['Name',  'Year',  'ROE',  'ROA',  'DE'])

newcolumns = (df.columns + [str(startyear)]).append(df.columns + [str(endyear)])

dfnew=pd.DataFrame(columns=newcolumns)

What I would like to have is (eg only years 2014 & 2015):

Name  ROE2014 ROA2014 DE2014 ROE2015 ROA2015 DE2015
AAPL  0.2     0.4     1.5    0.3     0.4     2.0
GOOGL 0.4     0.5     0.5    0.6     0.8     1.0

So far I only managed to get the new column names, but somehow I can't get my head around how to fill this new DataFrame.

Answer 1

Probably easier to create the new DataFrame, then adjust the column names:

# limit to data you want
dfnew = df[df.Year.isin(['2014', '2015'])]

# set index to 'Name' and pivot 'Year's into the columns 
dfnew = dfnew.set_index(['Name', 'Year']).unstack()

# sort the columns by year
dfnew = dfnew.sortlevel(1, axis=1)

# rename columns
dfnew.columns = ["".join(a) for a in dfnew.columns.values]

# put 'Name' back into columns
dfnew.reset_index()