Currently the process is quite painful. I have two or three historic data providers (Pinnacledata, CSIData, etc) and although I can convert from their format to NT's format easily, that's not sufficient. The problem is that NT mixes up all data from the same instrument in a, frankly, wrong way. My current solution, which is a nightmare, is to re-define the futures I am interested in for each individual provider and to use them as separate instruments. For instance I have: ZW (NT's default for wheat), CSI_ZW and PI_ZW. This way everything is kept separate, but ... I hope you see the problem.
My main desire is for NT to have a clear option to not mix the data for the *same* instrument under no circumstances, i.e. unless otherwise specified. My feeling is that this should be the default.
The second request is to allow in the namespace to easily distinguish between instruments from different data providers. One way to achieve that would be to allow provider specification in the name when referring to instruments and building lists, a.k.a. PI:ZW or CSI:ZW. Or even in the list name.
Finally, on the csv support: it should allow the data to not be cached in the database. Rather to be loaded every single time from the corresponding file. This way the files can transparently be updated by external programs. From my limited experience - multicharts does precisely that.
My understanding is that the extended csv support will come in the form of opening the APIs, which is fine with me, however, you might want to consider supporting more providers out of the box - for instance counting them as a form of a "broker" as well, thus, increasing the value of the multi broker license.
Comment