CODEWITHSUNDEEP
pythonpandaslistlambdaapply
Joe the Second
Joe the Second

Reputation: 362

How to apply a map in a lambda?

I have a column named "Increasing_sequence" in a dataset:

dict = {"Increasin_Sequence": [[[0.98, 1.1, 1.25], [1.18, 1.28]],[[1.2, 1.2], [1.1, 1.25]],[[0.85, 1.2, 1.29, 1.31, 1.4]],
[[1.19, 1.29, 1.39, 1.49]], [[1.0, 1.0, 1.0, 1.0]]] }
dt = pd.DataFrame(dict)


             Increasin_Sequence
0  [[0.98, 1.1, 1.25], [1.18, 1.28]]
1          [[1.0, 1.2], [1.1, 1.25]]
2     [[0.85, 1.2, 1.29, 1.31, 1.4]]
3         [[1.19, 1.29, 1.39, 1.49]]
4             [[1.0, 1.0, 1.0, 1.0]]

Each column consists of a list of "non-decreasing" lists. I want to keep lists which meet two following requirements (in a new column named "spikes"):

1- [the biggest number] - [the smallest number] >= .1

2- [the biggest number] > 1.2

so a desired output could be as following:

               spikes
0  [[0.98, 1.1, 1.25], [1.18, 1.28]]
1                      [[1.1, 1.25]]
2     [[0.85, 1.2, 1.29, 1.31, 1.4]]
3         [[1.19, 1.29, 1.39, 1.49]]
4                                 []

I have developed the following code:

   def spikes_finder(dt):
        dt['IncreasingSequences'].apply(lambda x: map(apply_spike_conditions, x))


   def apply_spike_conditions(increasing_sequence ):

        pick = increasing_sequence[-1]
        valley = increasing_sequence[0]
        pick_to_valley_difference = pick - valley
        if (pick_to_valley_difference >= .1) and (pick > 1.2):
            return increasing_sequence

Running this code, apply function doesn't execute, I also tried to use a for loop which is not a efficient way so I'd rather to use the apply or map function

Upvotes: 1

Views: 98

Answers (1)

Dani Mesejo
Dani Mesejo

Reputation: 61920

In general I find easier to put the steps into functions, so it be easier to understand, for example:

def conditions(lst):
    """This functions checks the filter conditions"""

    mi, *_, ma = lst  # extract the first and the last
    return ma > 1.2 and ma - mi > 0.1


def filter_cells(cell):
    """This function simply applies the filters"""
    return [lst for lst in cell if conditions(lst)]


dt['filtered'] = dt['Increasing_Sequence'].apply(filter_cells)
print(dt)

Output

                 Increasing_Sequence                           filtered
0  [[0.98, 1.1, 1.25], [1.18, 1.28]]  [[0.98, 1.1, 1.25], [1.18, 1.28]]
1          [[1.2, 1.2], [1.1, 1.25]]                      [[1.1, 1.25]]
2     [[0.85, 1.2, 1.29, 1.31, 1.4]]     [[0.85, 1.2, 1.29, 1.31, 1.4]]
3         [[1.19, 1.29, 1.39, 1.49]]         [[1.19, 1.29, 1.39, 1.49]]
4             [[1.0, 1.0, 1.0, 1.0]]                                 []

The notation

mi, *_, ma = lst

is known as extended iterable unpacking. It this context can be read as give me the first (mi), forget about the middle (*_) and also give me the last element (ma).

Regarding your functions, I think you are missing a return in spikes_finder, and perhaps is better to return True or False in apply_spike_conditions and use filter instead of map.

Note that in Python 3, both map and filter return iterables so you need to convert the result to a list.

Upvotes: 3

Related Questions