Yahoo Prices precision - pandas-datareader-0.9.0

[huvar]

Hello

Seems like yahoo is returning prices with more precision than it should. In the webpage they are displayed with tow decimals but in the html we can see extra decimals. The issue is that these extra decimals seem to be random.
Is there some way to fix it? Is it something that can be rounded safely? Is there a better free datasource?

https://finance.yahoo.com/quote/GOOGL/history

from pandas_datareader import data
start, end = '2018-01-11', '2021-01-11'
df = data.DataReader('GOOGL', start=start,end=end, data_source='yahoo')['Adj Close']

df
Date
2018-01-11 1112.050049
2018-01-12 1130.650024
2018-01-16 1130.699951
2018-01-17 1139.099976
2018-01-18 1135.969971
...
2021-01-05 1740.050049
2021-01-06 1722.880005
2021-01-07 1774.339966
2021-01-08 1797.829956
2021-01-11 1756.290039
Name: Adj Close, Length: 755, dtype: float64

[ikramersh]

This is caused by the way floating point numbers are stored in pandas dataframes.

There is a good discusion of the underlying cause in [https://github.com/pandas-dev/pandas/issues/16452] . It explains that it is a presentation issue and not an accuracy issue which means the numbers can safely be rounded. The issue is not caused by the data source but by the conversion into floating point numbers within pandas dataframes.

A simple solution is to set how pandas displays floating point numbers, for example if you want to display four decimal point precision use:

pd.options.display.float_format = '{:.4f}'.format
df = data.DataReader('GOOGL', start=start,end=end, data_source='yahoo')['Adj Close']

Note:

1. This will apply to the display of all floating point numbers in your pandas dataframes.
2. It does not affect the stored floating point values in the dataframe memory.

[huvar]

In this case it actually comes from the data source.
https://query1.finance.yahoo.com/v7/finance/download/GOOGL?period1=1515628800&period2=1516147200&interval=1d&events=history&includeAdjustedClose=true

[ikramersh]

Thanks for the data that shows the issue coming from the source.

The use of floating point numbers creates the extra decimals so Yahoo are probably using floating point storage to process the original prices before serving them up.