HDF5 for tick data with Python/pandas

quantstart · August 7th, 2013, 12:40 pm

Does anybody have experience/war stories to share using the above combination?In particular, on retail hardware, what can one expect in terms of performance, such as reads/writes per second to the file store?

quantstart · August 8th, 2013, 6:33 am

To clarify, I'm interested in using PyTables to read chunks of data into memory for analysis and/or to write data from multiple instruments. I'm keen on gauging a reasonable number of instruments that I'd be able to add to the file-store on a near real-time basis assuming access to an industry-grade tick feed.If anybody has done this then I'd love to hear from them!

Polter · August 8th, 2013, 11:10 am

There's been a thread on NP about this, with PyTables popping up:http://www.nuclearphynance.com/Show%20P ... Key=160458

quantstart · August 8th, 2013, 3:55 pm

Thanks Polter, I'll have a look.