CODEWITHSUNDEEP
pythonpandasdictionarydataframecategorical-data
user1111
user1111

Reputation: 331

Convert Categorical codes to Categorical values

I have a dataframe:

ga_deviceCategory_codes ga_channelgrouping_codes    ga_sourceMedium_codes   
        1.0                         6.0                      9.0
        1.0                         6.0                      9.0

Which I have converted into categorical codes from categorical values using :

data['ga_deviceCategory_codes'] = data['ga_deviceCategory'].astype('category').cat.codes
data['ga_channelgrouping_codes'] = data['ga_channelgrouping'].astype('category').cat.codes
data['ga_sourceMedium_codes'] = data['ga_sourceMedium'].astype('category').cat.codes

How do I get back to the original categorical values now from the above codes?

Upvotes: 3

Views: 6285

Answers (4)

kadee
kadee

Reputation: 8884

Assume you have a categorical series which you converted to a series of codes:

ser = df['col'].astype('category')
categorical = ser.dtype
ser_codes = ser.cat.codes

Then you can convert back to the original categorical series as follows:

pd.Series(pd.Categorical.from_codes(ser_codes, dtype=categorical), index=ser_codes.index)

Upvotes: 0

شیوا پاساراد
شیوا پاساراد

Reputation: 63

you could do this as follow:

df['col'] = df['col'].astype('category')
my_list = df['col'].cat.codes
my_values = df['col'].tolist()
result= dict(zip(my_list,my_values))

first convert your target columns to 'category' type, and then do label encoding on it using .cat.codes. then in line 3 of code, a list of target column real values is created, and in line 4 by zip, each values achived by cat.codes and related real values are consider as an element of a dict.

Upvotes: 1

jpp
jpp

Reputation: 164783

Category mappings are stored internally by Pandas, but not as a regular Python dictionary. You can create such a dictionary yourself to map backwards:

df['mycol'] = df['mycol'].astype('category')
d = dict(enumerate(df['mycol'].cat.categories))

Then map backwards:

df['mycol_codes'] = df['mycol'].cat.codes
df['mycol_reversed'] = df['mycol_codes'].map(d)

Be careful with this method. Make sure you create the dictionary straight after you convert to categories. When concatenating dataframes with categorical series, you may find the mapping changes.

Upvotes: 10

Andrew
Andrew

Reputation: 970

not 100% on what you are asking, but have you tried e.g.:

data['ga_sourceMedium_codes'].cat.categories

Upvotes: 0

Related Questions