bar df data type issues when converting

[sunjiawen]

In Entity.py,
You are converting the timestamp by multiply 1e9 and this cause the datatype to change to Float64 which is not accepted in the panda library and returns TypeError.

Can we do something like this?

if not df.empty:
                df.index = pd.to_datetime(
                    (df.index * 1e9).astype('int64'), utc=True,
                ).tz_convert(NY)

File "/Users/jiawensun/Alpaca/samplealgo/algo.py", line 59, in _get_prices
return barset.df
File "/usr/local/lib/python3.7/site-packages/alpaca_trade_api/entity.py", line 107, in df
df = bars.df.copy()
File "/usr/local/lib/python3.7/site-packages/alpaca_trade_api/entity.py", line 85, in df
df.index * 1e9, utc=True,
File "/usr/local/lib/python3.7/site-packages/pandas/core/tools/datetimes.py", line 603, in to_datetime
result = convert_listlike(arg, box, format)
File "/usr/local/lib/python3.7/site-packages/pandas/core/tools/datetimes.py", line 223, in _convert_listlike_datetimes
arg, _ = maybe_convert_dtype(arg, copy=False)
File "/usr/local/lib/python3.7/site-packages/pandas/core/arrays/datetimes.py", line 1914, in maybe_convert_dtype
data = data.astype(_NS_DTYPE)
File "/usr/local/lib/python3.7/site-packages/pandas/core/indexes/numeric.py", line 330, in astype
raise TypeError(msg)
TypeError: Cannot convert Float64Index to dtype datetime64[ns]; integer values are required for conversion

[thomasmarcel]

This looks like a decent solution. I'm having the same issue and, as you said, your Pull Request fixed it for me.

[ttt733]

Yes, this was caused by one of the breaking changes in pandas 0.24. I've merged your PR; thanks!