Datetime index issue?

[harshhacks]

I've recently been backtesting with a few CSV files and getting this warning everytime I run the backtest:

:19: UserWarning: Data index is not datetime. Assuming simple periods, but pd.DateTimeIndex is advised.
bt = Backtest(df_EURUSD, SMACrossover, cash=10000, commission=.000,exclusive_orders=True)

I have a time column which I have converted to DatetimeIndex and set as index but for some reason the backtesting engine adds an index column and uses that for backtesting which results in NaN values for any time-series based results like annual return.

Any ideas on where I might be going wrong?

[kernc]

The code uses:

backtesting.py/backtesting/backtesting.py

Lines 1047 to 1057 in e3cccdf

	# Convert index to datetime index
	if (not isinstance(data.index, pd.DatetimeIndex) and
	not isinstance(data.index, pd.RangeIndex) and
	# Numeric index with most large numbers
	(data.index.is_numeric() and
	(data.index > pd.Timestamp('1975').timestamp()).mean() > .8)):
	try:
	data.index = pd.to_datetime(data.index, infer_datetime_format=True)
	except ValueError:
	pass

backtesting.py/backtesting/backtesting.py

Lines 1079 to 1082 in e3cccdf

	if not isinstance(data.index, pd.DatetimeIndex):
	warnings.warn('Data index is not datetime. Assuming simple periods, '
	'but `pd.DateTimeIndex` is advised.',
	stacklevel=2)

Apparently, your index fails for:

assert isinstance(df_EURUSD.index, pd.DatetimeIndex), df_EURUSD.index

Can you show an example of your data frame?

[kernc]

Rename it to .txt.

[harshhacks]

aaares.txt

Attached the first 100 rows for your reference. Thanks a lot for taking a look!

[harshhacks]

For the strategy, I'm just doing something trivial like:

bt = Backtest(df, SMACrossover, cash=10000, commission=.002,exclusive_orders=True)

[kernc]

This works:

>>> df = pd.read_csv('aaares.txt', index_col='time', parse_dates=True)
>>> df.index.__class__
pandas.core.indexes.datetimes.DatetimeIndex

[harshhacks]

Thanks a lot! It's working great now.