Tukukkk

I've been trying to convert some of my functions where I use a for loop by using list comprehension. Here is my first version of the function,

def adstocked_advertising(data, adstock_rate):

    '''

    Transforming data with applying Adstock transformations

    data - > The dataframe that is being used to create Adstock variables

    adstock_rate -> The rate at of the adstock



    ex. data['Channel_Adstock'] = adstocked_advertising(data['Channel'], 0.5)

    '''

    adstocked_advertising = 

    for i in range(len(data)):

        if i == 0: 

            adstocked_advertising.append(data[i])

        else:

            adstocked_advertising.append(data[i] + adstock_rate * adstocked_advertising[i-1])            

    return adstocked_advertising

I want to convert it to this,

def adstocked_advertising_list(data, adstock_rate):

    adstocked_advertising = [data[i] if i == 0 else data[i] + adstock_rate * data[i-1] for i in range(len(data))]



    return adstocked_advertising

However, when viewing the df after running both functions I get two different values.

data['TV_adstock'] = adstocked_advertising_list(data['TV'], 0.5)

data['TV_adstock_2'] = adstocked_advertising(data['TV'], 0.5)

here is output,

data.head()

data.tail()

I am not too sure why the first two rows are the same and then from there the numbers are all different. I am new to list comprehension so I may be missing something here.

You need to refer to the previously generated element in the list, and list comprehensions are not well suited to this type of problem. They work well for operations that only need to look at a single element at once.
This question goes into more detail.

In your initial example, you use adstock_rate * adstocked_advertising[i-1]. The list comprehension version uses adstock_rate * data[i-1], which is why you are getting different results.

A standard for loop works just fine for your use case. You could switch to using enumerate, as for i in range(len(data)) is discouraged.

if data:

    res = [data[0]]

for index, item in enumerate(data[1:]):

        results.append(item + rate * data[index-1])

You've changed your logic in the list comp version. Originally, your else formula looked like:

data[i] + adstock_rate * adstocked_advertising[i-1]

But the list comprehension version looks like:

data[i] + adstock_rate * data[i-1]

The first version accesses the i-1th element of the result list, while the second version accesses the i-1th element of the input list.

index == 0 is only true once at the beginning of the list. Why not eliminate the conditional:

def adstocked_advertising(data, adstock_rate):

    if data:

        res = [data[0]]



        for i in range(1, len(data)):

            res.append(data[i] + adstock_rate * res[i-1])



        return res