I am not getting data for the expected time period. See code and output below

# region imports

from AlgorithmImports import *

from analyse import *

from datetime import datetime

# endregion

class NqTrend(QCAlgorithm):

def initialize(self):

self.set_start_date(2025, 6, 5)

self.set_cash(100000)

self.i = 0

self.future = self.add_future(

Futures.Indices.NASDAQ_100_E_MINI,

Resolution.MINUTE,

extended_market_hours=True,

data_normalization_mode=DataNormalizationMode.BACKWARDS_RATIO,

data_mapping_mode = DataMappingMode.LAST_TRADING_DAY,

contract_depth_offset = 0

)

self.future.set_filter(lambda future_filter_universe: future_filter_universe.front_month())

self.df = pd.DataFrame()

def OnEndOfAlgorithm(self):

self.debug(self.i)

self.debug(self.df)

def on_data(self, data: Slice):

self.i = self.i + 1

bar = data.bars.get(self.future.symbol)

if bar:

if bar.end_time.hour == 8 and bar.end_time.minute >= 31 and bar.end_time.minute <= 59:

self.df = pd.concat([self.df, pd.DataFrame({'time': bar.end_time, 'open': bar.open, 'high': bar.high, 'low': bar.low, 'close': bar.close, 'volume': bar.volume}, index=[1])], ignore_index=True)

elif bar.end_time.hour == 9 and bar.end_time.minute <= 30:

self.df = pd.concat([self.df, pd.DataFrame({'time': bar.end_time, 'open': bar.open, 'high': bar.high, 'low': bar.low, 'close': bar.close, 'volume': bar.volume}, index=[1])], ignore_index=True)

elif bar.end_time.hour == 9 and bar.end_time.minute >= 31 and bar.end_time.minute <= 59:

self.df = pd.concat([self.df, pd.DataFrame({'time': bar.end_time, 'open': bar.open, 'high': bar.high, 'low': bar.low, 'close': bar.close, 'volume': bar.volume}, index=[1])], ignore_index=True)

# Append rows Generate signals

elif (bar.end_time.hour >= 10 and bar.end_time.hour <= 12) or (bar.end_time.hour >= 13 and bar.end_time.hour <= 16):

self.df = pd.concat([self.df, pd.DataFrame({'time': bar.end_time, 'open': bar.open, 'high': bar.high, 'low': bar.low, 'close': bar.close, 'volume': bar.volume}, index=[1])], ignore_index=True)

# Close out if any positions (Assuming it crosses vwap or other stop loss); append rows

Here are the critical logic and configuration bugs in Sanjay's code that could cause only March data for a single day (not June) to appear, rather than the expected June slice:

1. Missing set_end_date in initialize

• The algorithm uses set_start_date(2025, 6, 5) but does not set an explicit end date. This means the backtest may by default only run for a single day (June 5) or potentially pick up a default or previously set global date (which could be March in some QC environments).
• Without set_end_date(2025, 6, 7), the algorithm will not collect a range of June data for the week as expected.

2. Time Zone Mismatch Risk

• The QCAlgorithm framework runs in New York time by default, but your research.ipynb uses qb.set_time_zone(TimeZones.HONG_KONG), which is a +12 hour difference.
• All your time-based filtering in on_data assumes a particular hour/minute that may not match the incoming bar timestamps if those are delivered in a different time zone than you expect.
• This can lead to nearly all bars being filtered out if you are, for example, looking for "9:00" bars but receiving "21:00" bars in UTC or another time, especially if your notebook and algorithm are compared directly.

3. Data Mapping Mode & Symbol Key

• In main.py, you use LAST_TRADING_DAY for data_mapping_mode, but in your research notebook, you use OPEN_INTEREST. This could result in the underlying contract being different for the same symbol identifier at a given time, depending on roll dates, which will produce subtly different minute data (especially if the front month changed between March and June).
• Also, data.bars.get(self.future.symbol) will only fetch data if the symbol exactly matches. If the filtered future.symbol corresponds to a contract that expired or didn't have data in June (but did in March), you will only see that March contract's bars.

4. Filtering on Bar Times

• Your on_data filtering logic is very restrictive, checking specific hour/minute combinations. If for any day, there is only a single bar that fits that criteria (and most are skipped due to an off time or timezone issue), your dataframe will look like only one day's worth of data, potentially for March if that's where a contract's last (or only) bar fell.
• If bar times are in UTC or another timezone, your lookups for "8:31 to 8:59", "9:00 to 9:30", etc. might never match, except in rare cases.

5. DataFrame Appending

• Each pd.concat([...], ignore_index=True) is correct and should append, but if bars almost never match your time criteria, you'll see just isolated rows—even if more data is available.

6. No Output Or Overwriting Bug, but Only Data Matches Filter

• Your self.df is not overwritten except with new rows, but since your filter is so narrow and potentially mismatched by timezone, only one matching row might appear across the whole backtest.

Summary of Bugs to Fix

• Add self.set_end_date(2025, 6, 7) in initialize to ensure the June range is covered.
• Review your time filtering logic and ensure it matches the bar time zone (likely NY by default in algorithms, but HONG_KONG in research).
• Unify data_mapping_mode between research and algorithm to avoid rolling contract mismatches.
• Consider logging every bar's time (self.debug(bar.end_time)) to confirm what times are actually available versus expected.
• Broaden or comment out the filtering temporarily to confirm data volume.

With these bugs, your algorithm is highly likely to only receive or keep the single bar which passes all filters on a previous contract's data, possibly leading to the "March data for one day only" issue instead of the intended June slice.

Sources: docs writing algorithms key concepts time modeling time zones, Stock Selection Strategy Based On Fundamental Factors, Future Option Multiple Contracts In Different Contract Months With Same Underlying Future Regression Algorithm, Seasonality Effect Based On Same Calendar Month Returns, Opening Range Breakout for Stocks in Play

After incorporating all the suggestions, I still do not see data from June. I only get 1-day of march data

class NqTrend(QCAlgorithm):

def initialize(self):

self.set_start_date(2025, 6, 1)

self.set_end_date(2025, 6, 6)

self.set_cash(100000)

self.i = 0

self.future = self.add_future(

Futures.Indices.SP_500_E_MINI,

Resolution.MINUTE,

extended_market_hours=True,

data_normalization_mode=DataNormalizationMode.BACKWARDS_RATIO,

data_mapping_mode = DataMappingMode.OPEN_INTEREST,

contract_depth_offset = 0

)

self.future.set_filter(lambda future_filter_universe: future_filter_universe.front_month())

self.df = pd.DataFrame()

def OnEndOfAlgorithm(self):

self.debug(self.i)

self.debug(self.df.iloc[-10:])

def on_data(self, data: Slice):

self.i = self.i + 1

bar = data.bars.get(self.future.symbol)

self.df = pd.concat([self.df, pd.DataFrame({'time': bar.end_time, 'open': bar.open, 'high': bar.high, 'low': bar.low, 'close': bar.close, 'volume': bar.volume}, index=[1])], ignore_index=True)

The out of sample period for free tier users is fixed to 3 months, this holds the recent data as unknown for you to test before going to live trading.