Ticker.history: history's index is not of type DateTimeIndex when dividend and split tables are empty

[ProgrammingDao]

Issue:

Ticker.history (line 295):

  df.index = df.index.tz_localize("UTC").tz_convert(data["chart"]["result"][0]["meta"]["exchangeTimezoneName"])

E AttributeError: 'Index' object has no attribute 'tz_localize'

Reason:

During the concatenation of a proper price table with empty dividend and split tables, the price table's index has changed from pandas.DatetimeIndex to pandas.Index because the empty tables had this index type.

Proposed solution:

Modify parse_actions (in utils.py) to initialise the dividend and split tables with :
dividends = _pd.DataFrame(columns=["Dividends"], index=_pd.DatetimeIndex([]))
splits = _pd.DataFrame(columns=["Stock Splits"], index=_pd.DatetimeIndex([]))

Local modules' versions:

finance : 0.1.69
pandas : 1.4.0
python : 3.10

[ProgrammingDao]

I have found out that the issue occurs when upgrading pandas from 1.3.5 to 1.4.0. Staying with pandas 1.3.5 doesn't cause the issue shared above.

[bma-diy]

Thank you for identifying the issue is with numpy. pip install pandas==1.3.5 resolved the issue for me.

Here are the error messages I experienced when using yfinance in a virtual environment:

  File "/home/ubuntu/myproject/env/lib/python3.8/site-packages/yfinance/base.py", line 295, in history
    df.index = df.index.tz_localize("UTC").tz_convert(
AttributeError: 'Index' object has no attribute 'tz_localize'

This happens whenever I try to download ticker data with 'period'.

[psbj]

Glad this was fixed, but the update isn't on pip, it's still at 0.1.69. I'm not familiar with how releases work on pip, is it automated and just takes a while or something else?

[ProgrammingDao]

@psbj you need to release a new version with the fix I think. There seems to be a GitHub action to do it. @ranaroussi tried to run it to deploy the newest package but it seems that it tried to deploy with version=0.1.169 that was already there instead of version=0.1.170.

[ProgrammingDao]

@psbj, @ranaroussi has just updated the version and released the version with the fix.

[keithtmccartney]

Thank-you so much, I had been bashing my head all weekend over a resolution when trying out Part Time Larry's (@hackingthemarkets) YouTube video "PyAlgoTrade Tutorial (Part 2) - Backtesting Basics" (https://www.youtube.com/watch?v=cdeftn1AZO8). New to Python, been a C# guy for years, this is a whole lot of fun!