Feed Pandas directly into Backtrader

[dfg12]

Hi,

I'd like to draw data from my database and feed it into backtrader via a Pandas dataframe, without the CSV intermediate step.

When I print my Pandas df, I get normal printed results.

Then I feed my df into my custom PandasData class

class PandasData(bt.feed.DataBase):
    lines = ('datetime', 'unix', 'open', 'high', 'low', 'close', 'volume')

    params = (
        ('datetime', None),
        ('unix', 'unix'),
        ('open', 'open'),
        ('high', 'high'),
        ('low', 'low'),
        ('close', 'close'),
        ('volume', 'volume')
    )

After which I add the data to bt:

   feed = PandasData(dataname=df)

    print('feed: ---', feed)

    cerebro.adddata(feed)

Once I add the strategy, however, I cannot access the data and cannot print it.

And get the following error message:

IndexError: array index out of range

Thank you in advance :))

[dfg12]

As an update, I am now just using the default PandasDataFeeds class like so:

    data = bt.feeds.PandasData(dataname=df)

    cerebro.adddata(data)

[run-out]

@dfg12 Could you share the top couple of rows from your dataframe?

[dfg12]

@run-out

Sure, can do!! :) Thanks for your response

datetime unix symbol open high low close volume exchange
0 2021-01-01 00:00:00 1609459200000 BTC/USDT 28923.63 28961.66 28913.12 28961.66 27.457032 binance
1 2021-01-01 00:01:00 1609459260000 BTC/USDT 28961.67 29017.50 28961.01 29009.91 58.477501 binance
2 2021-01-01 00:02:00 1609459320000 BTC/USDT 29009.54 29016.71 28973.58 28989.30 42.470329 binance
3 2021-01-01 00:03:00 1609459380000 BTC/USDT 28989.68 28999.85 28972.33 28982.69 30.360677 binance
4 2021-01-01 00:04:00 1609459440000 BTC/USDT 28982.67 28995.93 28971.80 28975.65 24.124339 binance

[run-out]

@dfg12 Try making sure datetime is a datetime object and not string. Move this to the index. You should be ok, but may need to identify the columns as per the docs.