Create columns from another column which is a list of items-CodePudding

Let's say that I have a DataFrame with column A which is a list of strings of the form "Type:Value" where Type can have 5 different values and Value can be anything. What I would like to do is to create new 5 columns (each having appropriate Type name) where the value in each column would be the list of items which has a given Type. So if I have (1 row for simplicity):

df = pd.DataFrame("A": [["Type1:Value1", "Type2:Value2", "Type1:Value3"]])

then the result should be:

df = pd.DataFrame("Type1": [["Value1", "Value3"]], "Type2":[["Value2"]])

CodePudding user response：

Goes without saying, but there is a probably a better way to do this.

import pandas as pd

df = pd.DataFrame({"A": [["Type1:Value1", "Type2:Value2", "Type1:Value3"]]})

buffer_dict = {}  # placeholder dict
for index, row in df.iterrows():
    for str_value in row['A']:
        str_list = str_value.split(':')
        key = str_list[0]  # these are just for readability
        value = str_list[1]
        buffer_dict.setdefault(key, []).append(value)  # set default to list and append values
buffer_dict.update((k, [v]) for k, v in buffer_dict.items())  # enclose values in list so we can convert to df
result = pd.DataFrame.from_dict(buffer_dict)
print(result)

Result:

              Type1     Type2
0  [Value1, Value3]  [Value2]

EDIT: I missed the part where there can only be 5 types. My solution was assuming this was unknown and will work for any amount of types.

CodePudding user response：

One Solution. This can be done on loop as well. But since the number of columns were small, the code is less automated.

df = pd.DataFrame({"A": [["Type1:Value1", "Type2:Value2", "Type1:Value3"]]})

df[['x','y','z']] = df.A[0]

df['type1'] = df.x.str.split(':').str[1]
df['type2'] = df.x.str.split(':').str[1]
df['type1'] = "["   df['type1']  ","  df.x.str.split(':').str[1]   "]"

# df['Type1'] = df.x.str.split(':').str[0]   ":"   "[["   df.x.str.split(':').str[1]   ","   df.z.str.split(':').str[1]   "]]"

print(df.drop(['A','x','y','z'], axis = 'columns'))



   type1   type2
0  [Value1,Value1]  Value1