How to create a dictionary for a 10,000 dataframe records and then access each dictionary record to make some calculations? - python-3.x

Question

I have a panda dataframe that has 10,000 records. The dataframe consists of 0 and 1 and looks like this:
C1 C2 C3 C4
0 0 1 1
0 1 0 0
1 0 1 1
My aim is to make each record as a dictionary which I assign a value for the dictionary for each column (each column has the same value):
C1 C2 C3 C4
{0: 10} {0: 11} {1: 15} {1: 13}
{0: 10} {1: 11} {0: 15} {0: 13}
{1: 10} {0: 11} {1: 15} {1: 13}
and then access the dictionary and make some calculation row by row and compare between the total of 0 and the total of 1. The number with the highest value will be added to a new column. For example for the first row 0 = 21 and 1 = 28 therefore 1 will be added to the new column. The second row 0 = 38 and 1 = 11 therefore 1 will be added:
C1 C2 C3 C4 New_column
{0: 10} {0: 11} {1: 15} {1: 13} 1
{0: 10} {1: 11} {0: 15} {0: 13} 0
{1: 10} {0: 11} {1: 15} {1: 13} 1

Answer 1

If you don't need the intermediate dicts, you can do some multiplications and sums:
values = [10, 11, 15, 13]
zeros = df.eq(0).mul(values).sum(axis=1)
ones = df.eq(1).mul(values).sum(axis=1)
df['New_column'] = ones.gt(zeros).astype(int)
# C1 C2 C3 C4 New_column
# 0 0 0 1 1 1
# 1 0 1 0 0 0
# 2 1 0 1 1 1
And if you do want the dicts, I would do it backwards and create the dicts after computing New_column:
df[['C1','C2','C3','C4']] = df.filter(like='C').apply(
lambda column: [{x:values[df.columns.get_loc(column.name)]} for x in column])
# C1 C2 C3 C4 New_column
# 0 {0: 10} {0: 11} {1: 15} {1: 13} 1
# 1 {0: 10} {1: 11} {0: 15} {0: 13} 0
# 2 {1: 10} {0: 11} {1: 15} {1: 13} 1