tickers.history fails when one ticker in list fails to download

[jakemdrew]

I have 1 of 564 tickers that causes tickers.history(period='30d', interval='1d') to fail. This is a valid ticker, but also fails download using yfinance:

[100%**] 1 of 1 completed

1 Failed download:

• KBSR: None

To Reproduce

tickers = Ticker(['KBSR'], asynchronous=True, progress=True) hist = tickers.history(period='30d', interval='1d')

Expected behavior

If a "failed to download" ticker is contained in the ticker_list, yahooquery will fail for the following internal error:

---

TypeError Traceback (most recent call last) in 4 5 # Get 30 days of price history for tickers ----> 6 hist = tickers.history(period='30d', interval='1d')

~.conda\envs\ML7331\lib\site-packages\yahooquery\ticker.py in history(self, period, interval, start, end, adj_timezone, adj_ohlc) 1240 else: 1241 data = self._get_data("chart", params) -> 1242 df = self._historical_data_to_dataframe(data, params, adj_timezone) 1243 if adj_ohlc and "adjclose" in df: 1244 df = self._adjust_ohlc(df)

~.conda\envs\ML7331\lib\site-packages\yahooquery\ticker.py in _historical_data_to_dataframe(self, data, params, adj_timezone) 1267 d = {} 1268 for symbol in self._symbols: -> 1269 if "timestamp" in data[symbol]: 1270 d[symbol] = _history_dataframe(data, symbol, params, adj_timezone) 1271 else:

TypeError: argument of type 'NoneType' is not iterable

Concern All valid data for 563 other tickers in my case are lost because of one failed download.

[danilogalisteu]

The same happens to me. The invalid tickers are not previously identified and removed with the "validate = True" argument. The data dictionary is returned from self._get_data with a NaN for each invalid symbol, which causes the reported error.

[danilogalisteu]

To sidestep the problem I modified the function _historical_data_to_dataframe as such:

    def _historical_data_to_dataframe(self, data, params, adj_timezone):
        d = {}
        invalid = []
        for symbol in self._symbols:
            if data[symbol] is not None:
                if "timestamp" in data[symbol]:
                    d[symbol] = _history_dataframe(data, symbol, params, adj_timezone)
                else:
                    d[symbol] = data[symbol]
            else:
                invalid.append(symbol)
        self._symbols = [s for s in self._symbols if not s in invalid]
        if self.invalid_symbols is not None:
            self.invalid_symbols += invalid
        else:
            self.invalid_symbols = invalid
        if all(isinstance(d[key], pd.DataFrame) for key in d):
            df = pd.concat(d, names=["symbol", "date"], sort=False)
            if "dividends" in df.columns:
                df["dividends"].fillna(0, inplace=True)
            if "splits" in df.columns:
                df["splits"].fillna(0, inplace=True)
            return df
        return d

[dpguthrie]

Is this still happening to folks? Not able to replicate using the example above.

https://www.loom.com/share/76783dfd46ae4e7dbc07797a6b323db2

[jakemdrew]

Resolved for me long ago

On Sat, Oct 15, 2022 at 1:32 PM Doug Guthrie @.***> wrote:

Is this still happening to folks? Not able to replicate using the example above.

https://www.loom.com/share/76783dfd46ae4e7dbc07797a6b323db2

— Reply to this email directly, view it on GitHub https://github.com/dpguthrie/yahooquery/issues/97#issuecomment-1279803040, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAQCS47RNHEH4WUB2C5HH4DWDL2B7ANCNFSM5HOFLRFA . You are receiving this because you authored the thread.Message ID: @.***>