The issue you're facing is common with MetaTrader 5's H1 timeframe data: when you run your script exactly on the hour, MT5 hasn't finished closing and generating the new hourly candlestick yet. It needs a few seconds to sync the latest tick data from the server and finalize the new bar. Here are several reliable solutions to fix this:

1. Add a Short Delay Before Fetching Data #

The simplest fix is to wait 10-15 seconds after the hour to give MT5 time to update its candlestick data. Use Python's time.sleep() for this:

import MetaTrader5 as mt5
import pandas as pd
import time

mt5.initialize()

# Wait 10 seconds to ensure MT5 completes the new H1 bar generation
time.sleep(10)

instrument = mt5.copy_rates_from_pos('BTCUSD', mt5.TIMEFRAME_H1, 0, 20)
instrument = pd.DataFrame(instrument)
instrument['time'] = pd.to_datetime(instrument['time'], unit='s')
instrument = instrument.set_index(['time'])

2. Trigger a Data Refresh by Fetching the Latest Tick #

MT5 will sync the latest data from the server when you request a tick for the symbol. This is a lightweight way to force a data update without waiting a fixed duration:

import MetaTrader5 as mt5
import pandas as pd

mt5.initialize()

# Fetch the latest tick to trigger a server data sync
latest_tick = mt5.symbol_info_tick('BTCUSD')
if latest_tick is None:
    print(f"Failed to fetch tick: Error code {mt5.last_error()}")
else:
    print(f"Synced latest tick price: {latest_tick.last}")

# Now fetch the H1 candlestick data
instrument = mt5.copy_rates_from_pos('BTCUSD', mt5.TIMEFRAME_H1, 0, 20)
instrument = pd.DataFrame(instrument)
instrument['time'] = pd.to_datetime(instrument['time'], unit='s')
instrument = instrument.set_index(['time'])

3. Use copy_rates_from with Explicit Time Bounds #

Instead of relying on position-based fetching, specify the end time as the current moment. This ensures you're requesting data up to the latest available bar:

import MetaTrader5 as mt5
import pandas as pd
from datetime import datetime

mt5.initialize()

# Fetch the last 20 H1 bars ending at the current time
end_time = datetime.now()
instrument = mt5.copy_rates_from('BTCUSD', mt5.TIMEFRAME_H1, end_time, 20)
instrument = pd.DataFrame(instrument)
instrument['time'] = pd.to_datetime(instrument['time'], unit='s')
instrument = instrument.set_index(['time'])

4. Smart Retry: Check for the Latest Bar Time #

For the most robust solution, verify if the last fetched bar matches the current hour. If not, wait and retry until you get the updated data:

import MetaTrader5 as mt5
import pandas as pd
import time
from datetime import datetime

mt5.initialize()

def get_updated_h1_data(symbol, bar_count=20, max_retries=5, wait_interval=5):
    for attempt in range(max_retries):
        rates = mt5.copy_rates_from_pos(symbol, mt5.TIMEFRAME_H1, 0, bar_count)
        if rates is None:
            print(f"Attempt {attempt+1}: Failed to fetch data. Retrying...")
            time.sleep(wait_interval)
            continue
        
        df = pd.DataFrame(rates)
        df['time'] = pd.to_datetime(df['time'], unit='s')
        df = df.set_index(['time'])
        
        # Get the current hour (rounded down to the start of the hour)
        current_hour = datetime.now().replace(minute=0, second=0, microsecond=0)
        
        # Check if the latest bar is from the current hour
        if df.index[-1] >= current_hour:
            return df
        
        print(f"Attempt {attempt+1}: Latest bar is {df.index[-1]}. Waiting for update...")
        time.sleep(wait_interval)
    
    raise RuntimeError("Failed to retrieve updated H1 data after multiple attempts")

# Fetch the data using the smart retry function
instrument = get_updated_h1_data('BTCUSD', 20)
print(instrument['open'].tail(5))

Each method has its tradeoffs:

• Fixed delays are simple but may wait longer than necessary.
• Tick-based refresh is lightweight but doesn't guarantee the bar is ready immediately.
• Time-bound fetching is reliable but requires handling time zone consistency with MT5.
• Smart retry is the most robust, as it only waits until the bar is actually available.

内容的提问来源于stack exchange，提问作者GuyS