Yes, JDB is much newer than KDB, so it might not be as refined yet as KDB could be, and of course I will report back any feeedback I can give... but my question was aimed at anybody who might already have made some comparisons, who are actually using or have used, or tested, both systems. Especially anybody who works with tick data.
For me it is important that it can handle large datasets without producing any limit errors or bottlenecks where KDB would not. I have no idea right now how big a dataset gets before KDB would struggle, I have only recently started to learn it. Anyway, I was wondering if there are any people out there who use Q/KDB and have tried or are using J (and possibly have tried out the new JDB) to see how they compare in terms of performance and capability. There are many things in q which are exactly the same as in J but go under a different name. I also note that J feels more self consistent and complete as a language. On the other hand, the dictionaries and keyed table support in q/KDB are extermely useful. It would be interesting to compile a list of how to "translate" q/KDB into J/JDB to acheive similar tasks, and to compare the performance and beauty of expression. Thanks, Matthew. On Tue, Sep 30, 2008 at 3:08 PM, Devon McCormick <[EMAIL PROTECTED]> wrote: > Matthew - as far as I understand it, JDB is _much_ newer than KDB. A > disadvantage of this is that KDB almost certainly has better raw performance > as they've been working on this problem for quite a while. In fact, someone > like you who works with both systems would be in a better position to tell > everyone here about relative performance. > > An advantage of JDB's newness is that the early users of it can influence > its direction. Personally, I'm more interested in ease of use, stability, > integration with J (and, potentially, other environments), and ability to > handle large datasets. Performance, assuming it's adequate, would be less > of a consideration for me than these other things. > > I'm very interested in trying out JDB's capabilities but have not yet done > so. Please report your experiences as you become more familiar with it. > > Regards, > > Devon > > On Tue, Sep 30, 2008 at 5:03 AM, Matthew Brand <[EMAIL PROTECTED]>wrote: > >> Hi All, >> >> I am learning q/KDB to use at work currently, but am more familiar >> with J. Does anybody know of a good essay which compares the >> performance and capabilities of q/KDB with J? >> >> The biggest differences I can see so far is the built-in KDB side and >> the built in use of dictionaries and tables in q - which is very >> useful. JDB may be trying to provide a similar functionality to KDB? I >> prefer J at the moment because it feels like it fits together better, >> but that may be because it is what I am used to using. >> >> Has anybody done a performance/functionality comparison between q/KDB >> and J/JDB? Is anybody using J to process tick data? >> >> >> Thanks, >> Matthew. >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm >> > > > > -- > Devon McCormick, CFA > ^me^ at acm. > org is my > preferred e-mail > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
