I am pleased to announce that the RayforceDB columnar database, developed by Lynx Capital Partners, is now an open source project.
RayforceDB is an implementation of the array programming language Rayfall (in the same way that kdb+ is an implementation of k/q), which inherits the ideas embodied in k and q. However, RayforceDB uses Lisp-like syntax, which, as our experience has shown, significantly lowers the entry threshold for beginners and also makes the code much more readable and easier to maintain. However, the implementation of K syntax remains an option for enthusiasts of this type of notation.
RayforceDB is written in pure C with a minimum of external dependencies, the executable file size does not exceed 1 megabyte on all platforms (tested and actively used on Linux, macOS, and Windows), and the executable file is the only thing you need to deploy to get a working instance. Additionally, it is possible to compile to Webassembly and run in a browser. However, in this case, automatic vectorization is not available.
RayforceDB was developed by a company that provides infrastructure for the most liquid financial markets. As you might expect, the company has extremely high requirements for data processing speed. The effectiveness of the tool can be determined by visiting the following link: https://rayforcedb.com/content/benchmarks/bench.html
The connection with the Python ecosystem is facilitated by an external library, which is available here: https://raypy.rayforcedb.com
RayforceDB offers all the features that users of columnar databases would expect from modern software of this kind. Please find the necessary documentation and a link to the project's GitHub page at the following address: https://rayforcedb.com
owenm•2mo ago
This is super interesting - aside from the benchmarks are there any good examples of loading and querying data from external sources? Any plans to support other formats like parquet?
viktorsovietov•2mo ago
So far, we haven't had any particular need for native support for parquet files. This type of import can be done satisfactorily via RayPy, and there is also a plugin that allows you to access kdb+ instances.
viktorsovietov•2mo ago
I am pleased to announce that the RayforceDB columnar database, developed by Lynx Capital Partners, is now an open source project.
RayforceDB is an implementation of the array programming language Rayfall (in the same way that kdb+ is an implementation of k/q), which inherits the ideas embodied in k and q. However, RayforceDB uses Lisp-like syntax, which, as our experience has shown, significantly lowers the entry threshold for beginners and also makes the code much more readable and easier to maintain. However, the implementation of K syntax remains an option for enthusiasts of this type of notation.
RayforceDB is written in pure C with a minimum of external dependencies, the executable file size does not exceed 1 megabyte on all platforms (tested and actively used on Linux, macOS, and Windows), and the executable file is the only thing you need to deploy to get a working instance. Additionally, it is possible to compile to Webassembly and run in a browser. However, in this case, automatic vectorization is not available.
RayforceDB was developed by a company that provides infrastructure for the most liquid financial markets. As you might expect, the company has extremely high requirements for data processing speed. The effectiveness of the tool can be determined by visiting the following link: https://rayforcedb.com/content/benchmarks/bench.html
The connection with the Python ecosystem is facilitated by an external library, which is available here: https://raypy.rayforcedb.com
RayforceDB offers all the features that users of columnar databases would expect from modern software of this kind. Please find the necessary documentation and a link to the project's GitHub page at the following address: https://rayforcedb.com