How to recode/ map shared columns for dataframe stored in a dictionary?-CodePudding

I want to recode the 'Flavor' field, which both data sets share.

I successfully stored the data as data frames in a dictionary, but the names assigned (for ex. 'df_Mike') are strings and not callable/ alterable objects.

Do let me know where I'm going wrong and explain why.

name = ['Mike', 'Sue']

d = {}

for n in name:
    url = f'https://raw.githubusercontent.com/steflangehennig/info4120/main/data/{n}.csv'
    response = urllib.request.urlopen(url)
    d[n] = pd.DataFrame(pd.read_csv(response))

flavor = {1: 'Choclate', 2: 'Vanilla', 3:'Mixed'}

for df in d:
    df.map({'Flavor': flavor}, inplace = True)

Error code:

      1 flavor = {1: 'Choclate', 2: 'Vanilla', 3:'Mixed'}
      3 for df in d:
----> 4     df.map({'Flavor': flavor}, inplace = True)

AttributeError: 'str' object has no attribute 'map'

CodePudding user response：

You are iterating over the keys in the dictionary. If you want to iterate over the values, you can use dict.values(). For example:

for df in d.values():
    df.map({'Flavor': flavor}, inplace = True)

CodePudding user response：

It would be better to iterate over the dict items. So, use

for name, df in d.items():

CodePudding user response：

You can try

for n in name:
    url = f'https://raw.githubusercontent.com/steflangehennig/info4120/main/data/{n}.csv'
    df = pd.read_csv(url)
    d[n] = df

for df in d.values():
    df['Flavor'] = df['Flavor'].map(flavor)

or only in the first loop

for n in name:
    url = f'https://raw.githubusercontent.com/steflangehennig/info4120/main/data/{n}.csv'
    df = pd.read_csv(url)
    df['Flavor'] = df['Flavor'].map(flavor)
    d[n] = df