How do I select based on a string another pandas dataframe in python?

Question

I have one main dataframe and several FX-Rate dataframes. The main one is loaded with data and with one column, which tells you the corresponding FX-Rate Table.

If i read one row in the main table, it tells me at the end: FX_TABLE_2 This comes as a string. How do I open than the FX_TABLE_2, which is a dataframe object?

eg

FX_TABLE_1 = pd.read_excel(path6, 'FX_TABLE_1',index_col=0)
FX_TABLE_2 = pd.read_excel(path6, 'FX_TABLE_2',index_col=0)
FX_TABLE_3 = pd.read_excel(path6, 'FX_TABLE_3',index_col=0)
FX_TABLE_4 = pd.read_excel(path6, 'FX_TABLE_4',index_col=0)
FX_TABLE_5 = pd.read_excel(path6, 'FX_TABLE_5',index_col=0)
FX_TABLE_6 = pd.read_excel(path6, 'FX_TABLE_6',index_col=0)
FX_TABLE_7 = pd.read_excel(path6, 'FX_TABLE_7',index_col=0)


for ind in Mainframe.index:
    ###### select FX Table for row
    FX_Table = Mainframe['FX_Table'][ind]
    print(type(FX_Table))
    print(type(FX_TABLE_5))

so i get a str for the first element, but based on this str out of the FX_Table column, i want to choose the correct dataframe. If i print the second element from the code, it is a <class 'pandas.core.frame.DataFrame'> and cannot be opened with a <class 'str'>.

Could anyone help? would be really nice!

It is basically select a dataframe from a output of a formula which is a str-element

Answer 1

for instance this is the mainframe, but much more easier:

main = {
  "EXP": [1, 2, 3, 4],
  "EXP2": [1, 2, 3, 4],
  "FX_Table": ["FX_TABLE_5", "FX_TABLE_2","FX_TABLE_3","FX_TABLE_4"],
    "Currency": ["EUR", "EUR", "CNY", "CNY"]
}

#load data into a DataFrame object:
main = pd.DataFrame(main)

the fx-rates table for one year is also attached as image fx_rates_table_example

so i want to read the mainframe, get the value of the column, eg FX_TABLE_2 --> then open the correct FX_TABLE-Dataframe like attached.

at the moment a get a STR-Element, but cannot open with this the correct dataframe, because it is another type.

Answer 2

maybe this picture or structure of the process makes it more clear

process image

Answer 3

Ok, so it looks like you have quite a misconception here.

Please note, you cannot load dataframe into another as they are separate objects, however, you can load data from one df into another in different ways, using concat , merge and etc.

small example for you below, I hope it will lead you forward, it has comments of each step:

main = {
  "EXP": [1, 2, 3, 4],
  "EXP2": [1, 2, 3, 4],
  "FX_Table": ["FX_TABLE_1", "FX_TABLE_2","FX_TABLE_1","FX_TABLE_2"],
    "Currency": ["EUR", "EUR", "CNY", "GBP"]
}

#load data into a DataFrame object:
main = pd.DataFrame(main)

path6 = r'C:\Sandbox\test.xlsx'
Excel_File = pd.ExcelFile(path6)  # reads whole excelFile at once
dfs_dict = {sh:Excel_File.parse(sh) for sh in Excel_File.sheet_names} # loading each sheet to DataFrame and each dataframe to dictionary
print(dfs_dict.keys()) # will show each sheet name loaded
print(dfs_dict['FX_TABLE_2'])  # will show data from single DataFrame by it's key in the dictionary (which is the same as sheet name)

#adding new column to each DataFrame saved in the dictionary:
for key in dfs_dict.keys():
 dfs_dict[key]['Sheet_Name'] = key

print(dfs_dict['FX_TABLE_2']) # will show DataFrame with new column which can be used as unique ID

# now you can add data from other tables into Main or New_main dataframe, - up to you. I will use Sheet Name and Currencty as unique ID for merging
New_main = pd.merge(main, dfs_dict['FX_TABLE_1'], how='outer', left_on=['FX_Table', 'Currency'], right_on=['Sheet_Name','currency'])

print(New_main)

   EXP  EXP2    FX_Table Currency currency  ...  2008  2009  2010  2011  Sheet_Name
0    1     1  FX_TABLE_1      EUR      EUR  ...   3.0  3.05   3.1  3.15  FX_TABLE_1
1    2     2  FX_TABLE_2      EUR      NaN  ...   NaN   NaN   NaN   NaN         NaN
2    3     3  FX_TABLE_1      CNY      CNY  ...   2.0  2.05   2.1  2.15  FX_TABLE_1
3    4     4  FX_TABLE_2      GBP      NaN  ...   NaN   NaN   NaN   NaN         NaN

or as per your comment:

FX_Tab_list = main['FX_Table'].to_list # load table names to the list from Main DF and loop the list
for table in FX_Tab_list:
    dfs_dict[table] = your code lands here
    #for example:
    dfs_dict[table]['difference'] = dfs_dict[table]['2009'] - dfs_dict[table]['2008']