Native support of Pandas DataFrame?
DaveSkender opened this issue · comments
We've seen several Python users get frustrated over issues related to converting quotes from Pandas DataFrame into iterable quotes. Are there opportunities to add a native strongly typed DataFrame interface implementation to the .NET library? I'm assuming this might then enable better interface design for our Python implementation.
Some paths to explore:
- improve existing Python
stock-indicators
package code to be more resilient or to provide enhanced error handling / messages - improve documentation website by either / or:
- a dedicated page for Pandas DataFrame users
- add full downloadable example usage code, like the Stock Indicators for .NET | Examples
- the mentioned native implementation of .NET DataFrame
For now, to help users, I've started a specific support discussion on the topic, including a list of reported examples.
cc: @LeeDongGeon1996
Looks like 2
is almost done :)
About 1
, how do you think about providing helper method for converting into Iterable[Quote]?
About
1
, how do you think about providing helper method for converting into Iterable[Quote]?
I think what you've done with Exception handling might be enough; unless you can think of some better utility method that'd work better. Right now they have to do some custom transform with a number of methods using zip()
or numpy.vectorize()
or iterrows()
, and to do it correctly with all necessary data type conversions and considerations, with code like:
quotes_list = [
Quote(d,o,h,l,c,v)
for d,o,h,l,c,v
in zip(df['date'], df['open'], df['high'], df['low'], df['close'], df['volume'])
]
A utility would have more validation and error handling and be in a somewhat simpler interface, like:
quotes = df.to_quotes('date', 'open', 'high', 'low', 'close', 'volume');
or
quotes = df_to_quotes(df['date'], df['open'], df['high'], df['low'], df['close'], df['volume']);
For this issue, I’m not sure where to go at this point. The example usage we’ve given so far somewhat explains how to use a DataFrame, maybe that’s enough?
@DaveSkender
I also think that what we've done for dealing with DataFrame
is enough. Most of cases were the problem of what the beginners are usually struggled with (except for locale
issue). So now let's just see how it goes.