Erro ao tentar juntar dois arrays do numpy que tem mesmo número de linhas mas com número de colunas diferentes

O array features2 é (5074382, 82) dtype('float64') numpy.ndarray

features_encoded : (5074382, 9276434) dtype('float64') scipy.sparse.csr.csr_matrix

features_final = np.column_stack((features2, features_encoded))

A linha acima gera o erro:

ValueError: all the input array dimensions for the concatenation axis must match exactly, but along dimension 0, the array at index 0 has size 5074382 and the array at index 1 has size 1

Eu preciso justar os dois arrays acima. Como fazer?

O dataset que estou utilizando tem algumas colunas que sao categorias. Nelas apliquei o OneHotEncoder. Depois, tentei juntas o array com as features numericas e o array que saiu do OneHotEncoder, formando, assim, um único array com todas as features!

Segue o código completo:

import numpy as np
    import pandas as pd
    import matplotlib.pyplot as plt
    import seaborn as sns
    pd.options.display.max_columns = None  #Display all dataframe columns in a Jupyter Python Notebook
    pd.set_option('display.max_rows', 1000)
    get_ipython().run_line_magic('matplotlib', 'inline')

    CIC2019 = pd.read_csv(r"DrDoS_DNS.csv")

    remove =lambda x:x.strip()# remove the blancks in columns names

    columns = list(CIC2019.columns)

    new_columns =list(map(lambda x:x.strip(),columns))# removing blamcks

    CIC2019 = pd.read_csv(r"DrDoS_DNS.csv", names =new_columns, header = None, skiprows=1,nrows=None)
    CIC2019.rename(columns={"Unnamed: 0": "ID"}, inplace=True)

    CIC2019 = CIC2019.dropna()
    CIC2019.isna().sum()

    features = CIC2019.drop("Label", axis =1)

    # # Handling categorical attributes

    from sklearn.preprocessing import OneHotEncoder
    encoder = OneHotEncoder()

    CIC2019["Label"]

    Label_encoded = encoder.fit_transform(CIC2019["Label"].to_numpy().reshape(1,-1))

    features[["Flow ID","Source IP","Timestamp","SimillarHTTP","Destination IP"]]

    features2 = features.drop(["Flow ID","Source IP","Timestamp","Destination IP","SimillarHTTP"], axis =1)
    features2 = features2.to_numpy()

    features_encoded = encoder.fit_transform(features[["Flow ID","Source IP","Timestamp","Destination IP",]].to_numpy())
    #"SimillarHTTP" : error when you added this
    # # Training - Linear Regression

    features_final = np.column_stack((features2, features_encoded))