Python Pandas：如何根据主框架中的列值映射数据框字典中的值？

Question

I am trying to create a column in my main dataframe DF_MAIN by mapping over values from another dataframe that is part of a larger dictionary of dataframes DF_DICTDF. The dataframe from which to map in DF_DICTDF depends on the value in a column in DF_MAIN.

For greater detail, I am creating a "RATE" column in the main dataframe and each dataframe in the dictionary corresponds to a table of currency exchange rates for each month. The rates need to be mapped over based on a "CURRENCY" column in DF_MAIN. The dictionary has a naming convention as follows: 2018January, 2018February, 2018March... which correspond to values in the "YEARMONTH" column of DF_MAIN.

DF_MAIN is one dataframe that contains columns "NAMES", "PRICES", "CURRENCIES" (ie CNY, USD, JPY, EUR...) and "YEARMONTH" (ie 2018January, 2018February...).

DF_DICTDF is a dictionary of dataframes. The names of the dataframes within this dictionary are simply YEARMONTH (ie 2018January, etc.) so that each dataframe contains currency cross rates corresponding to that month. The column and row index for each YEARMONTH dataframe is simply currency symbols (ie CNY, USD, JPY, EUR, etc.).

I tried to apply the following:

DF_MAIN['RATE']=DF_MAIN['CURRENCY'].map(DF_DICTDF[DF_MAIN['YEARMONTH']].set_index('CURRENCY')['EUR'].to_dict())

When I try to run this, I get the following error:

TypeError: 'Series' objects are mutable, thus they cannot be hashed

---

Same as above but breaking it out into two steps for clarity:

### Create dictionary of cross rates to map
crossrates=DF_DICTDF[DF_MAIN['YEARMONTH']].set_index('CURRENCY')['EUR'].to_dict()

### Final Mapping of Values:
DF_MAIN['RATE']=DF_MAIN['CURRENCY'].map(crossrates)

All help appreciated. If this is a duplicate please provide a link.

Answer 1

I was able to make it work using the following:

def convert(row):
    x='EUR' 
    y=row['CURRENCY']
    z=row['YEARMONTH']
    return DF_DICTDF[z].set_index('CURRENCY').loc[x,y]

DF_MAIN['RATE']=DF_MAIN.apply(lambda row: convert(row),axis=1)

(Will edit at a later time to elaborate).