How to use a collection as an element of a Python array?

Question

I am developing a project using scikit-Learn (and pandas to handle the data) to predict the results of football matches based on previous results. As a project methodology, for each row of the dataset used in the prediction, the result is determined using the last three games of the teams, which I put in an array of Numpy. I got an error when I used to try to add the set with the last three games to the pandas dataset that will be used in the classification. The code:

import pandas as pd
import numpy as np
from sklearn.metrics import confusion_matrix
from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeClassifier

ds_resultados = pd.read_csv("E0_new2.csv")

ds_resultados["Date"] = pd.to_datetime(ds_resultados.Date)

data = ds_resultados[["FTHG", "FTAG", "HTHG", "HTAG", "HS", "AS", "HST", "AST"]]
target = ds_resultados["FTR"]

#CÓDIGO PARA PEGAR OS TRÊS ÚLTIMOS RESULTADOS PARA CADA JOGO

cont = 0;

for index, row in ds_resultados.iterrows():
    auxHome = ds_resultados[((ds_resultados["HomeTeam"] == row["HomeTeam"])
    | (ds_resultados["AwayTeam"] == row["HomeTeam"]))]
    auxHome = auxHome[auxHome["Date"] < row["Date"]].sort_values(by="Date", ascending=True).head(n =                 3)

listaTarget = np.array([]).astype('int64')

if auxHome.shape[0] == 3:
    for i in range(0, 3):
        linha = auxHome.iloc[i]

        listaTarget = np.append(listaTarget, linha["FTR"])

    data = data.append(row[["FTHG", "FTAG", "HTHG", "HTAG", "HS", "AS", "HST", "AST"]])

    target = target.append(pd.Series([]), ignore_index = True)
    target.at[cont] = listaTarget #linha que, quando executada, gera o erro
    cont = cont + 1

When the line

target.at[cont] = listaTarget

is runs, I get the following error

ValueError: setting an array element with a sequence.

If necessary, I leave below the data set (E0_new2.csv) used with appropriate subtitles (the "FTR" column is used as a variable target, I put it in the dataset for explanatory purposes only):

   print(data.head(n=10))
      FTHG  FTAG  HTHG  HTAG  HS  AS  HST  AST  FTR
   0     2     1     1     0   8  13    6    4    1
   1     2     0     1     0  12  10    4    1    1
   2     0     2     0     1  15  10    6    9   -1
   3     0     3     0     2   6  13    1    4   -1
   4     1     2     1     2  15  15    2    5   -1
   5     2     0     1     0  19   6    5    0    1
   6     2     2     1     1  11   6    4    5    0
   7     0     2     0     1   9  17    3    8   -1
   8     4     0     2     0  18   5    8    2    1
   9     0     0     0     0  18  16    3    6    0

Legend of the data: FTHG (Full-Time Home Team Goals): how many goals did the home team score in the match FTAG (Full-Time Away Team Goals): how many goals did the visiting team score in the match HTHG (Half-Time Home Team Goals): How many goals did the home team score until halftime HTAG (Half-Time Away Team Goals): How many goals did the visiting team score until halftime HS (Home Team Shots): how many shots the home team gave to the opponent’s goal AS (Away Team Shots): how many shots the visiting team gave to the opponent’s goal HST (Home Team Shots on Target): how many shots did the home team hit AST (Away Team Shots on Target): how many shots did the visiting team hit FTR (Full-Time Result): final result of the match, with 0 indicating draw, 1 indicating victory of the home team and -1 indicating victory of the visiting team

How could I form a collection in which one of the elements is another collection for that purpose?

Answer 1

On the line:

target.at[cont] = listaTarget

The variable target is Series type. When you call the .[at cont], you are referencing the element in position cont of the Series which is a unique value and not a list or array. It’s happening that you’re trying to add a list/array/Series into a variable that only fits a single element.

To solve this you can create a Dataframe variable and use append/Concat to add new elements to it.