Why all values are the same in nested dictionary?-CodePudding

I'm trying to generate a nested dictionary but when for loops finish all the values are the same in every key.

def frequency(df, headers, numerical, target):
di = {}
if len(numerical) == 0:
    print("No NM")
else:
    for i in numerical:
        if i in headers:
            headers.remove(i)

    for i in range(len(headers)):
        attdic = {}
        valdic = {}
        aux_list = df[headers[i]].to_list()
        aux_list = list(OrderedDict.fromkeys(aux_list))
        tar_list = df[target].to_list()
        tar_list = list(OrderedDict.fromkeys(tar_list))

        for j in range(len(aux_list)):
            print(aux_list[j])

            for k in range(len(tar_list)):
                temp_df = df.loc[(df[headers[i]] == aux_list[j]) & (df[target] == tar_list[k])]
                tempvar = len(temp_df.index)   1  

                valdic[tar_list[k]] = tempvar
                attdic[aux_list[j]] = valdic
                di[headers[i]] = attdic

Output:

{'Outlook': {'overcast': {'No': 2, 'Yes': 5}, 'rainy': {'No': 2, 'Yes': 5}, 'sunny': {'No': 2, 'Yes': 5}}, 'Windy': {False: {'No': 3, 'Yes': 5}, True: {'No': 3, 'Yes': 5}}}

Expected output:

{'Outlook': {'overcast': {'No': 3, 'Yes': 3}, 'rainy': {'No': 3, 'Yes': 4}, 'sunny': {'No': 2, 'Yes': 5}}, 'Windy': {False: {'No': 4, 'Yes': 6}, True: {'No': 3, 'Yes': 5}}}

CodePudding user response：

Notice that you create one valdic for each i, but you should have created one for each (i,j), so you can move it inside the loop for j, and maybe it works

def frequency(df, headers, numerical, target):
di = {}
if len(numerical) == 0:
    print("No NM")
else:
    for i in numerical:
        if i in headers:
            headers.remove(i)

    for i in range(len(headers)):
        attdic = {}
        aux_list = df[headers[i]].to_list()
        aux_list = list(OrderedDict.fromkeys(aux_list))
        tar_list = df[target].to_list()
        tar_list = list(OrderedDict.fromkeys(tar_list))

        for j in range(len(aux_list)):
            valdic = {} # create a new valdic for each j
            print(aux_list[j])

            for k in range(len(tar_list)):
                temp_df = df.loc[(df[headers[i]] == aux_list[j]) & (df[target] == tar_list[k])]
                tempvar = len(temp_df.index)   1  

                valdic[tar_list[k]] = tempvar
            attdic[aux_list[j]] = valdic
        di[headers[i]] = attdic