CODEWITHSUNDEEP
pythonrfunctionpandasmethods
ooolllooo
ooolllooo

Reputation: 351

Method like argument in function

I want use method in python / pandas like argument in a function. For example rolling statistics for dataframe:

def rolling (df, prefix = 'r', window = 3, method = 'here I wanna choose a method' ):
    for name in df.columns:
        df[prefix + name] = df[name].rolling(window).'here this method been called'
    return df

'mean()' or 'sum()' or whatever ... like 

df.rolling(2).sum()

I worked 95% time in R, and in R it's simple (put function as an argument or return any function ). But in python I noob. So I creating package to make things easier for me. Like:

def head(x,k = 3):
    return x.head(k)

What function in python help me to use method like argument in a function?

#some data
import numpy as np
import pandas as pd
from pandas_datareader.data import DataReader
from datetime import datetime
ibm = DataReader('IBM',  'yahoo', datetime(2000,1,1), datetime(2016,1,1))

ibm2 = rolling(ibm,'rr', 5, 'sum') # something like this

Upvotes: 1

Views: 106

Answers (3)

chepner
chepner

Reputation: 531165

A method is an attribute like any other (it just happens to be callable when bound to an object), so you can use getattr. (A default value of None is nonsense, of course, but I didn't want to reorder your signature to make method occur earlier without a default value.)

def rolling (df, prefix='r', window=3, method=None):
    for name in df.columns:
        obj = df[name].rolling(window)
        m = getattr(obj, method)
        df[prefix + name] = m()
    return df

Upvotes: 0

piRSquared
piRSquared

Reputation: 294258

I do this

def rolling (df, prefix='r', window=3, method='method_name'):
    for name in df.columns:
        df[prefix + name] = df[name].rolling(window).__getattribute__(method)()
    return df

Upvotes: 1

Artyer
Artyer

Reputation: 40811

You can use getattr with a str of the name of the method. This gets the attribute with that name from the object (In this case, a method)

def rolling (df, prefix='r', window=3, method='sum'):
    for name in df.columns:
        df[prefix + name] = getattr(df[name].rolling(window), method)()
    return df

Or you could just pass in the method. When calling it, the first argument will be self.

def rolling (df, prefix='r', window=3, method=DataReader.sum):
    for name in df.columns:
        df[prefix + name] = method(df[name].rolling(window))
    return df

Upvotes: 3

Related Questions