I understand your point, but it’s not really necessary to “blur the line”. I mean you can save the Reuters data in the database as they are and then leave the user the option to “normalise” them in the query result (or manage a second db with normalised data to speed up the queries).
Except for the annual data, the interim data (which are the most interessant, because more recent) are practically unusable because of a lot of different periods lenghts and units.
For example in my NYSE db for Income Statement there are 13 differents periods!
f = io.StringIO()
download_reuters_financials([“RTLR”],f, universes=[“nyse”], interim=True)
financials = pd.read_csv(f, parse_dates=[“SourceDate”, “FiscalPeriodEndDate”])
(financials[‘StatementPeriodLength’].map(str) + financials[‘StatementPeriodUnit’]).unique()
array([‘12W’, ‘16W’, ‘3M’, ‘6M’, ‘13W’, ‘12M’, ‘14W’, ‘5M’, ‘4M’, ‘1M’,
‘52W’, ‘2M’, ‘9M’], dtype=object)
For the Cash Flow Statement even worse with 23 different periods!:
['16W', '28W', '40W', '52W', '12M', '3M', '6M', '9M', '26W', '39W', '13W', '12W', '24W', '36W', '53W', '5M', '4M', '7M', '8M', '27W', '10M', '2M', '15M']
Actually is very difficult to perform a simple query like for example “give me the last 4 quarterly result for Total Revenue”.
I think the point should be addressed, because it could be the “added value” provided by Quantrocket.
As a programmer I’ve already written a Java Service that download the Reuters data from IB in XML format and put them in a relational database but the difficult task (and therefore the “added value”) would be in the post-process/homogenization of data. That “added value” would contribute to justify a paid subscription for QuantRocket.
Another use case is the following one: I test my algo in Quantopian because of their free long fundamentals data history and then trade it live with QuantRocket. Actually it’s not possible without a lot of work, because the Quantopian data are quarterly and yours (in the simplest case) annually.
QuantRocket looks very promising to me and I hope you will consider my request constructively.
We can also ask to the other members of the QuantRocket community, which is their opinion about this point.