On 1 May 2009 at 16:45, Hae Kyung Im wrote:
| it may be slightly off topic but I was wondering if any of you heard
| about using netCDF format (or similar) to handle tick data?
I know of places that use hdf5 so it likely that someone may also be using
| I thought kdb would be a nice option but the price seems a bit too
| high for my purpose. Do you know of any good open source alternative?
| Also is there any package to connect R with kdb?
Yes, you can get it off the (public access, as I recall) kx.com website. I
looked at it for a few days---and even enhanced the existing R / kdb package
with corrected support for sub-second time types between R and Kx with a
patch you find on my blog at http://dirk.eddelbuettel.com/blog---but we
decided to go with a competing product (for which I've since written an
internal R package connecting to their C++ API). If you want to evaluate Kx,
you get a free-as-in-beer 32 bit binary that will run for two hours after
which you need to relaunch.
The is a lot of cool stuff listed at the bottom of the 'column-oriented DBMS'
page on Wikipedia. Someone should nudge these towards open-source tick data
bases. Jeff and I talked about it but alas no free time...
Three out of two people have difficulties with fractions.
> it may be slightly off topic but I was wondering if any of you heard
> about using netCDF format (or similar) to handle tick data?
> I thought kdb would be a nice option but the price seems a bit too
> high for my purpose. Do you know of any good open source alternative?
> Also is there any package to connect R with kdb?
If you want to have a database primarily for analysis, testing etc.:
32-bit kdb+/q prior version v.3.3 (and after 2.8 I think) is free for
production use (if you managed to download the copy from kx site in the
Note that those versions are no longer available for download. Latest
32-bit versions are still free for proof-of-concept and development use.
On Tue, Feb 2, 2016 at 6:00 PM, Daniel Cegiełka <[hidden email]>