[pandas] make custom time-series data

import pandas as pd
import dask.dataframe as dd
import numpy as np
import os
from sklearn.datasets import make_classification

# parameters
n_samples = 2000
start = pd.to_datetime('2017-01-01')
end = pd.to_datetime('2017-01-31')

def random_datetimes_or_dates(start, end, out_format='datetime', n=10):
    start_u = start.value//10**9
    end_u = end.value//10**9

    return pd.DatetimeIndex((10**9*np.random.randint(start_u, end_u, n, dtype=np.int64)).view('M8[ns]'))

# features
df = make_classification(n_samples = n_samples, n_features = 10)
df_raw = pd.DataFrame(df[0], columns = ['var1', 'var2', 'var3', 'var4', 'var5', 'var6', 'var7', 'var8', 'var9', 'var10'])
df_raw['var1'] = np.random.randint(2, size=n_samples) + 1
df_raw['var2'] = np.random.randint(100000, size=n_samples)
df_raw['var3'] = np.random.randint(1000000, size=n_samples)
df_raw['trx_inRow'] = 1

# timestamps
df_timestamp = pd.DataFrame({'date': random_datetimes_or_dates(start, end, out_format= 'datetime', n=n_samples).sort_values()})
df_raw['timestamp'] = pd.to_datetime(df_timestamp['date'], format='%Y-%m-%d', errors='ignore')

# class
df_raw['class'] = df[1]

del df, df_timestamp