ES future question

Here’s an improved approach for fetching ES (S&P 500 E-mini Future) historical daily data in QuantBook that better matches exchange official OHLC and removes spurious bars returned for holidays like Juneteenth:

import numpy as np
import pandas as pd
from AlgorithmImports import *

# 1) Add the ES (S&P 500 E-mini) future at daily resolution
qb = QuantBook()
es_security = qb.add_future(Futures.Indices.SP_500_E_MINI, Resolution.DAILY)
es_symbol = es_security.symbol

# 2) Fetch 30 days of daily history, with fill_forward and extended_market_hours both False
es_history = qb.history(
    [es_symbol],
    30,
    Resolution.DAILY,
    fill_forward=False,
    extended_market_hours=False
)

# Exit early if no data
if es_history.empty:
    print("No ES data returned.")

else:
    # If multi-index, select just the ES data
    if isinstance(es_history.index, pd.MultiIndex):
        es_history = es_history.xs(es_symbol, level="Symbol")
        if isinstance(es_history, pd.Series):
            es_history = es_history.to_frame()

    # Ensure index is datetime
    es_history.index = pd.to_datetime(es_history.index)

    # 3) Remove rows for known US holidays (e.g., Juneteenth, Independence Day, etc.)
    us_holidays = [
        pd.Timestamp("2023-06-19"),  # Juneteenth example
        pd.Timestamp("2023-07-04"),  # Independence Day example
        # Add other holidays as needed
    ]
    mask = np.isin(es_history.index.date, [h.date() for h in us_holidays])
    es_history = es_history[~mask]

    # 4) Display the cleaned DataFrame
    print(es_history)

Key points:

• fill_forward=False and extended_market_hours=False provide the cleanest official session bars, minimizing artificial fills or after-hours influences.
• Rows for holidays are explicitly dropped using a list of US holiday dates. You can extend us_holidays as needed.
• Any “ghost” bar on a non-trading day should be filtered out this way, so your results match what you’d see on IBKR or TradingView.

Let me know if you need to automate the holiday-list (using Python's holidays library or QuantConnect's trading calendars)!

Sources: Pre Holiday Effect, Upcoming Holiday Momentum for Amazon, Futures Fast Trend Following, with Trend Strength, Key Concepts - QuantConnect.com, Futures - QuantConnect.com

Using fill_forward=False and extended_market_hours=False doesn't change anything.
So QuantConnect doesn't exclude holidays by default and I have to manually remove them?

I just consolidated the hourly data to daily OHLC for a single day. QuantConnect's data is consistent while Trading View is not. The hourly data for QuantConnect is consistent with Trading View's. But something odd is happening when Trading View consolidates high resolution data into Daily OHLC, hence the discrepancy. Still not sure why QuantConnect has data for 06/19 for ES though as it's a US holiday.