Set decimal precision of a pandas dataframe column with a datatype of Decimal

Question

I have a pandas dataframe with two columns, col 1 with text in it and col 2 with decimal values.

Key	Value
A	1.2089
B	5.6718
B	7.3084

I use the '.apply' function to set the data type of the value column to Decimal (Python Decimal library). Once I do this the Value column goes from a 4 decimal place value to 43 decimal places. I have attempted to use the .getcontect.prec = 4 to no avail.

The data frame is constructed from reading a CSV file with the same format as the table above. All the decimal numbers in the value column are only given to 4 decimal places.

import pandas as pd
from decimal import *

def get_df(table_filepath):
    df = pd.read_csv(table_filepath)
    getcontect.prec = 4
    df['Value'] = df['Value'].apply(Decimal)

The above code is what I have tried but still results in a output with the value column values having 43 decimal places rather than the 4 decimal places each value should have as read from the csv file.

The result I get when I print the dataframe is:

Key	Value
A	1.20890000000003046807250939309597015380859375
B	5.67180000000000318323145620524883270263671875
B	7.30838399999999969077180139720439910888671875

I only want 4 decimals of precision because these values will be use to do some maths later on and I want to work with the exact values I provided.

Answer 1

This can be modified by changing the print options for floats, however it will modify how every float datatype is printed

pd.set_option('display.float_format', '{:.10f}'.format)

Keep in mind that this is only the way it's printed. The value is stored in the dataframe, with every decimal.

On the other hand, you can restrict decimals by:

df.Value = df.Value.round(4)

But this will round depending the fifth decimal. Last option would be to use np.ceil or np.floor but since this wont support decimals, an approach with multiplication and division is requierd:

precision = 4
df['Value_ceil'] = np.ceil(df.Value * 10**precision) / (10**precision)
df['Value_floor'] = np.floor(df.Value * 10**precision) / (10**precision)

Answer 2

Fixed the issue, seems to be related to how Decimal converts from float to decimal. Setting the Values column to be of data type string then converting to Decimal got me the result I desired.

def get_df(table_filepath):
    df = pd.read_csv(table_filepath)
    df['Value'] = df['Value'].apply(str) 
    df['Value'] = df['Value'].apply(Decimal)

Key	Value
A	1.2089
B	5.6718
B	7.3084