I have an option of either maintaining 50 different pandas dataframes having 5 columns as shown below with index (date)
df1.head(1) date