How to lookup two different ranges based on row's value Pandas dataframe

Question

I'm trying to do a conditional vlookup but with pandas. Here's the data i'm using

n_age_scores

type n     aging_n     mini_n   percent_n
new        <30 days       0       0.5543
new        31-50 days     31      0.6446
new        51-100 days    51      0.3134

e_age_scores

type e           aging_e      mini_e   percent_e
expansion        <30 days       0       0.33543
expansion        31-50 days     31      0.4446
expansion        51-100 days    51      0.6134

Dataframe

type        age    score
new          33
new          12
expansion    3
new          4
expansion    100

What I want to do is populate score with the percent column of either dataframe based on if the row type is new or expansion, an approximate match of value percent.

How do I do this with Pandas?

n_age_scores = aging_score_mapping.iloc[:,0:4] 
e_age_scores = aging_score_mapping.iloc[:,-4:9]


    if df['deal_type'] == 'Expansion':
       df = merge.e_age_scores(df, on='age_score')
       if df['deal_type'] == 'new':
          df = merge.n_age_scores(df, on='age_score')

I'm not sure how to do this but I think i need to loop and merge with a approximate match and populate age_score with percent_n depending on the type.

Is this possible with pandas?

Answer 1

You can define your custom function to extract the data from the two dataframes and use it with apply .

If I understood correctly what you want, the code below shoud do the job.
I called age_map the resulting dataframe (those with the score column to be populated and collector the function which extract the data from n_age_scores and e_age_scores .

def collector(row):
    if row['type'] == 'new':
        return n_age_scores.loc[n_age_scores['mini_n'] < row['age']].iloc[-1]['percent_n']
    elif row['type'] == 'expansion':
        return e_age_scores.loc[e_age_scores['mini_e'] < row['age']].iloc[-1]['percent_e']

age_map['score'] = age_map.apply(collector, axis=1)

Using the sample dataframes you provide, age_map is:

        type  age    score
0        new   33  0.64460
1        new   12  0.55430
2  expansion    3  0.33543
3        new    4  0.55430
4  expansion  100  0.61340