Proper syntax of a loop in python

Question

I am new in python and I would like to see the solution from experts on the below, as choosing values from some dataframe based on values from some other one, is something that I am using often.

I have the first df below.

data = {'Date': ['31/03/1947', '30/06/1947', '30/09/1947', '31/12/1947', '31/03/1948', '30/06/1948', '30/09/1948', '31/12/1948', '31/03/1949', '30/06/1949'],
        'Italian GDP': [-0.057750, -0.054695, -0.052334, -0.051142, -0.050267, -0.049659, -0.048984, -0.048839, -0.046507, -0.045941],
        'US GDP': [-0.011017, -0.008948, -0.007276, -0.006526, -0.005046, -0.002011, -0.001592, -0.000720, 0.000085, 0.000334],
        'EA GDP': [0.009119, 0.010925, 0.011530, 0.014639, 0.015634, 0.016439, 0.018998, 0.025592, 0.032806, 0.035710],
        'FR GDP': [-0.011773, -0.010264, -0.009310, -0.009126, -0.006450, -0.005746, -0.004998, -0.004780, 0.001206, 0.004616],
        'DE GDP': [-0.030926, -0.023653, -0.023418, -0.021585, -0.010145, -0.006971, -0.005937, -0.005850, -0.005215, -0.000496],
        'CZ GDP': [-0.017287, -0.013185, -0.011872, -0.006948, -0.004398, -0.004028, -0.001028, -0.000814, 0.000349, 0.001409],
        'HU GDP': [-0.068923, -0.060180, -0.058478, -0.057246, -0.053871, -0.053105, -0.052404, -0.052222, -0.050352, -0.049721],
        'UK GDP': [-0.029143, -0.028303, -0.027973, -0.027784, -0.025187, -0.024922, -0.024092, -0.022788, -0.022478, -0.021911],
        'NL GDP': [-0.011381, -0.010251, -0.009614, -0.008191, -0.007078, -0.006006, -0.005573, -0.000319, -0.000199, 0.000804],
        'RO GDP': [-0.082121, -0.076776, -0.074611, -0.073923, -0.070768, -0.060687, -0.060526, -0.054234, -0.047550, -0.032161]}

# read in with 
df = pd.DataFrame(data).set_index('Date')

and a second df below:

Values = {'GDPs': ['Italian GDP', 'US GDP', 'EA GDP', 'FR GDP', 'DE GDP', 'CZ GDP', 'HU GDP', 'UK GDP', 'NL GDP', 'RO GDP'],
        'Observations': [89, 281, 89, 169, 105, 85, 89, 169, 85, 89],
        'Round Down': [1.0, 5.0, 1.0, 3.0, 2.0, 1.0, 1.0, 3.0, 1.0, 1.0],
        'Round Up': [2.0, 6.0, 2.0, 4.0, 3.0, 2.0, 2.0, 4.0, 2.0, 2.0]}
# read in with 
roundposnew = pd.DataFrame(Values).set_index('GDPs')

For the second df, I would like to add two additional columns that would give me the corresponding "Round Up" and "Round Down" values of the first df. For example, I would like to get the first value and second value of the "Italian GDP", the 5th and 6th value of the "US GDP" etc. I have written the code below just for the "Round Down" values (one of the columns i want to add) but is not working, it gives just the result for last loop value. Can you please advise what would be the proper way of doing it? Thank you in advance!

for i in df.columns:
    fal = df.columns.get_loc(i)
    ld5 = df.iloc[int(roundposnew.loc[i,'Round Down']-1),int(fal)]
ld5

>>> -0.08212129722076356

Answer 1

• When roundposnew is created, Round Down and Round UP should be int , not float (eg 1 vs. 1.0 ) because the value is being used to index another dataframe.
  • Then there's no reason for converting to int in the loop
• Write the values to a list and then create the new column
  • Dataframe columns should be created all at once with a list or array of values.
  • Trying to create a column in a loop, as your were probably doing, repeatedly assigns each value from the loop, to all the rows in the column.
• As a note, the logic of what you're doing, manually selecting these particular values, doesn't make sense. It seems like there is something needing a programmatic implementation.
  • What is the logic behind selecting the values from df to be added to roundposnew ? Why those specific rows, for each column?

import pandas as pd

Values = {'GDPs': ['Italian GDP', 'US GDP', 'EA GDP', 'FR GDP', 'DE GDP', 'CZ GDP', 'HU GDP', 'UK GDP', 'NL GDP', 'RO GDP'],
          'Observations': [89, 281, 89, 169, 105, 85, 89, 169, 85, 89],
          'Round Down': [1, 5, 1, 3, 2, 1, 1, 3, 1, 1],
          'Round Up': [2, 6, 2, 4, 3, 2, 2, 4, 2, 2]}

# read in with 
roundposnew = pd.DataFrame(Values).set_index('GDPs')

# round up and down list
ru = list()
rd = list()

# loop to add values to lists
for i in df.columns:
    fal = df.columns.get_loc(i)  # this is an int, doesn't need int(fal)
    rd.append(df.iloc[roundposnew.loc[i,'Round Down'] - 1, fal])
    ru.append(df.iloc[roundposnew.loc[i, 'Round Up'] + 1, fal])

# add lists to dataframe
roundposnew['rd'] = rd
roundposnew['ru'] = ru

             Observations  Round Down  Round Up        rd        ru
GDPs                                                               
Italian GDP            89           1         2 -0.057750 -0.051142
US GDP                281           5         6 -0.005046 -0.000720
EA GDP                 89           1         2  0.009119  0.014639
FR GDP                169           3         4 -0.009310 -0.005746
DE GDP                105           2         3 -0.023653 -0.010145
CZ GDP                 85           1         2 -0.017287 -0.006948
HU GDP                 89           1         2 -0.068923 -0.057246
UK GDP                169           3         4 -0.027973 -0.024922
NL GDP                 85           1         2 -0.011381 -0.008191
RO GDP                 89           1         2 -0.082121 -0.073923

Answer 2

I assume you want to add 2 additional columns, you could do it like this: first you add 2 columns filled with 0 values (or whatever you want) and later you update these values with your rounddown/roundup formula

roundposnew['Rounded up'] = 0.
roundposnew['Rounded down'] = 0.
for i in df.columns:
    fal = df.columns.get_loc(i)
    ld5 = df.iloc[int(roundposnew.loc[i,'Round Down']-1),int(fal)]
    ld6 = df.iloc[int(roundposnew.loc[i, 'Round Up'] + 1), int(fal)]
    roundposnew.loc[i,'Rounded down'] = ld5
    roundposnew.loc[i, 'Rounded up'] = ld6
    print('values for '+i+' are: \n rounded down: '+str(ld5)+' \n rounded up: '+str(ld6))
print(roundposnew)

Output looks like:

             Observations  Round Down  Round Up  Rounded up  Rounded down
GDPs                                                                     
Italian GDP            89         1.0       2.0   -0.051142     -0.057750
US GDP                281         5.0       6.0   -0.000720     -0.005046
EA GDP                 89         1.0       2.0    0.014639      0.009119
FR GDP                169         3.0       4.0   -0.005746     -0.009310
DE GDP                105         2.0       3.0   -0.010145     -0.023653
CZ GDP                 85         1.0       2.0   -0.006948     -0.017287
HU GDP                 89         1.0       2.0   -0.057246     -0.068923
UK GDP                169         3.0       4.0   -0.024922     -0.027973
NL GDP                 85         1.0       2.0   -0.008191     -0.011381
RO GDP                 89         1.0       2.0   -0.073923     -0.082121

Hope this shows you a way to solve your issue!