Hello,
I was a member of the old group and an active user of ib_insync, RIP Ewald.? I have used the IB API using different approaches, I give some more background below, and prefers ib_insync by far.? I do have some help routines that I may contribute but I am wondering if it is best to try to incorporate them into the library itself or simply as notebook/recipe.? Any advice on that appreciated.
About the discussion on storage, I currently do not use any form of database, but store bars on disk using parquet, possibly out of laziness.? I have thought about moving to some database for a while, and it would probably be some SQL based variant.? I am also not yet looking at dask level, not even close, everything done on a fast PC running Windows.
On a side topic, I much prefer the groups.io format compared to either discord or the github discussion.? Thank you for starting this.
Background on my IB API usage
- started with R and the IBrokers package, this was pre pandas days, so I was using R for data processing and python for everything else.? R is nice for table data but a pain for scripting (file and string manipulation)
- moved to a single language approach after numpy matured.? I started by using IbPy, this was before the introduction of the official python library by IB
- moved to the official python library when it become available, did not really like it, also tried C++ and liked it even less
- discovered ib_insync (tws_async at that time) and never looked back, this is paired with pandas for all data processing currently
- recently discovered polars, and somehow I prefer the polars approach to pandas, faster and somehow a more consistent syntax.? Not ready to switch yet as the library is not as rich as pandas yet.
Background on myself
- engineer working in the semiconductor industry
- strong interest in trading, including automated trading
- not very active right now, but looking into volatility based trading
- living in California but originally from Belgium
