How Do I Create Multiple Dataframes That Are Named Based On The Value Of Another Dataframe?

I have a dataframe with about 2.5M rows and would like to split this up into about 10000 rows (one for each unique value of TICKER), and to have the name be df(i) for each value of

Solution 1:

Yes, you can do with groupby:

for k,d in df.groupby('ticker'):
    print(d)
    print('-'*30)

Output:

  ticker        date assets
0      A  04-30-2020     40
1      A  07-30-2020     50
6      A  10-31-2020    100
------------------------------
  ticker        date assets
2     BB  05-31-2020     60
3     BB  08-31-2020     70
------------------------------
  ticker        date assets
4    CCC  06-30-2020     80
5    CCC  09-30-2020     90
------------------------------

Solution 2:

I think you need this,

import pandas as pd

df = pd.DataFrame({"ticker": ["A", "A", "BB", "BB", "CCC", "CCC", "A"], "date": ["04-30-2020", "07-30-2020", "05-31-2020", "08-31-2020", "06-30-2020", "09-30-2020", "10-31-2020"], "assets": ["40","50","60","70","80","90","100"]})

import sys

thismodule = sys.modules[name]

ls = df.ticker.unique().tolist()

for i in ls: setattr(thismodule, "df"+i, df[df.ticker==i])

display(dfA,dfBB,dfCCC)

Output

Note:

• If you need to refer to those df names in a loop, please use the following:

for i in ls: print(eval("df"+i))

• name is enclosed within double underscores on both sides(underscore underscore name underscore underscore)