I tried
method 1

## only show part of code here  
def initialize(context):  
    global stock  
    stock = sid(8554) # We will be trading SPY  
    algo.schedule_function(  
            rebalance,  
            date_rules.every_day(),  
            algo.time_rules.market_open(),  
        )    

def rebalance(context, data) :  
    prices = data.history(stock, 'close', 9+1, '1d')  
    print('price={}'.format(prices))

and get result 1 (partial result)

2004-03-18 09:31  PRINT price=  
2004-03-05 00:00:00+00:00    116.29  
2004-03-08 00:00:00+00:00    115.26  
2004-03-09 00:00:00+00:00    114.66  
2004-03-10 00:00:00+00:00    112.98  
2004-03-11 00:00:00+00:00    111.30  
2004-03-12 00:00:00+00:00    112.69  
2004-03-15 00:00:00+00:00    111.08  
2004-03-16 00:00:00+00:00    111.73  
2004-03-17 00:00:00+00:00    113.00  
2004-03-18 00:00:00+00:00    112.65  
Freq: C, Name: Equity(8554 [SPY]), dtype: float64

2004-03-19 09:31  PRINT price=  
2004-03-08 00:00:00+00:00    114.857  
2004-03-09 00:00:00+00:00    114.259  
2004-03-10 00:00:00+00:00    112.585  
2004-03-11 00:00:00+00:00    110.911  
2004-03-12 00:00:00+00:00    112.296  
2004-03-15 00:00:00+00:00    110.692  
2004-03-16 00:00:00+00:00    111.339  
2004-03-17 00:00:00+00:00    112.605  
2004-03-18 00:00:00+00:00    112.505  
2004-03-19 00:00:00+00:00    112.380  
Freq: C, Name: Equity(8554 [SPY]), dtype: float64  

Here we can see that the close price at 2004-03-17 00:00:00+00:00
changed from 113.00 to 112.605

I also tried to use method 2 with minute data

def handle_data(context, data):  
    today = get_datetime().date()  
    # Do nothing unless the date has changed  
    if today == context.date:  
        return  
    # Set the new date  
    context.date = today

    ## daily historical close price  
    prices = data.history(stock, 'close', 9+1, '1d')  
    print('price={}'.format(prices))  

But the result 2 was the same

I then tried method 3 pipeline

stocks = symbols('SPY')

def initialize(context):  
    global stock  
    stock = sid(8554) # We will be trading SPY and tlt  
    context.i = 0 # Set index for windows  
    context.prices = []  
    # Attach the pipeline defined in my_pipe so we have data to use  
    attach_pipeline(make_pipeline(context), name='my_data')

    # Rebalance every day, right after market open.  
    algo.schedule_function(  
        rebalance,  
        date_rules.every_day(),  
        algo.time_rules.market_open(),  
        # algo.time_rules.market_close(hours = 1),  
    )    


def make_pipeline(context):  
    universe = Filters.StaticAssets(stocks)  
    price = USEquityPricing.close.latest  
    pipe = Pipeline(  
        columns={  
            'price': price,  
        },  
        screen=universe  
    )  
    return pipe


def before_trading_start(context, data):  
    # Store our pipeline output DataFrame in context.  
    context.output = pipeline_output('my_data')  
    context.prices.append(*context.output.price.values)


def rebalance(context, data):  
    # Skip first 9 days to get full windows  
    context.i += 1  
    # print('i={}'.format(context.i))  
    if context.i < 9:  
        return  
    price = context.prices[-1]  
    print('price={}'.format(context.prices))  

and I got result 3

2004-03-17 09:31  PRINT price=[116.05, 116.29000000000001, 115.26000000000001, 114.66, 112.98, 111.3, 112.69, 111.08, 111.73]  
2004-03-18 09:31  PRINT price=[116.29000000000001, 115.26000000000001, 114.66, 112.98, 111.3, 112.69, 111.08, 111.73, 113.0]  

Since I append the latest close price to a list, it won't happen a changing issue like data.hitory

The last value 113.0 in the list at 2004-03-18 09:31 is the close price at 2004-03-17, which is the same as the method 1 before the data.history changed its value

There are also another dates having the same issue around 1/24/2016 and 4/2/2017
From this result 3, I get different trading results since it doesn't have the same close price as method 1 on certain dates.

What's happened to the change?
Is there anything wrong with my code?