Question: Is there an efficient way to get the currency?

[ymyke]

The history() function doesn't return the currency the prices are in.

In order to get the currency, one has to invoke the get_info() function, which is rather costly – on the order of a few seconds.

Is there a more efficient way to get the currency?

[ValueRaider]

Possible but slightly hacky. Yahoo returns currency with price data but no official way to add it to Pandas table. So question is where to put it?

Can add a .currency attribute to table but it'll disappear as soon as you modify table, e.g.:

df = yf.download(...)
print(df.currency)  # Good
print(df)  # Pandas won't show currency
df = df.copy()
print(df.currency)  # AttributeError exception

Or add a new currency property to Ticker - will persist but won't help download().

[ymyke]

Thanks! I tried this, however, it doesn't work:

df = yf.download("MSFT")
print(df.currency)  # produces AttributeError

Am I missing something?

[ymyke]

I saw that Yahoo returns some meta data. Example below. The currency is in there. Along with all kinds of other information that could be useful here and there.

Feature request: Would be great if – upon retrieval – you could simply store the raw "meta" dictionary in the Ticker object so it would be accessible if necessary/useful.

(Maybe this is more or less what you were suggesting above and what you were describing there is not how it works today but how it could work in the future?)

{'currency': 'USD',
 'symbol': 'COIN',
 'exchangeName': 'NMS',
 'instrumentType': 'EQUITY',
 'firstTradeDate': 1618407000,
 'regularMarketTime': 1669237204,
 'gmtoffset': -18000,
 'timezone': 'EST',
 'exchangeTimezoneName': 'America/New_York',
 'regularMarketPrice': 45.57,
 'chartPreviousClose': 43.39,
 'previousClose': 43.39,
 'scale': 3,
 'priceHint': 2,
 'currentTradingPeriod': {'pre': {'timezone': 'EST',
   'start': 1669194000,
   'end': 1669213800,
   'gmtoffset': -18000},
  'regular': {'timezone': 'EST',
   'start': 1669213800,
   'end': 1669237200,
   'gmtoffset': -18000},
  'post': {'timezone': 'EST',
   'start': 1669237200,
   'end': 1669251600,
   'gmtoffset': -18000}},
 'tradingPeriods': [[{'timezone': 'EST',
    'start': 1669213800,
    'end': 1669237200,
    'gmtoffset': -18000}]],
 'dataGranularity': '1m',
 'range': '1d',
 'validRanges': ['1d', '5d', '1mo', '3mo', '6mo', '1y', '2y', 'ytd', 'max']}

[ValueRaider]

Oh it's not implemented, I'm just saying it could be but need to think best way.

If you have good idea then submit a PR to dev

[fredrik-corneliusson]

Seems like pandas does support adding extra properties to dataframes https://pandas.pydata.org/pandas-docs/stable/development/extending.html#define-original-properties

However it might be be better to use @ymyke suggestion and store it on the Ticker instead.

[ymyke]

Alright, here's the PR, @ValueRaider: https://github.com/ranaroussi/yfinance/pull/1198 – Turned out to be very easy.

[ymyke]

Note re download: It wouldn't be straightforward to add the functionality to the download method. Adding the info to the underlying dataframe object is not a solution since you can have several tickers with differing meta info (e.g., currencies) in one download call. So the information would have to be added to the dataframe itself, which would be a major interface change and maybe not advisable for the time being.

This means that both ways to retrieve several tickers simultaneously (yf.download and yf.Tickers) do not expose the meta info at this point.