TRIADB is discontinued, page is maintained here for the history.
Introduction
TriaClick is my long standing effort to revitalize Relational and Topic Maps data model. Associative filtering, similar to Qlik associative engine, has been implemented for the first time with a fast columnar DBMS on non-volatile memory and tested with a relatively large file on commodity hardware.
Screencast
Screen Capture Demo of TriaClick, a python library that implements associative, semiotic, hypergraph technology on top of ClickHouse columnar DBMS and MariaDB. We show the execution of commands from two python console applications that are built with TriaClick library. The various operations (methods) of our Chain Query Language (CQL) aim to make the processing pipeline of data integration, exploratory data analysis and visualization easier, faster, intuitive, and more efficient and accurate for the database/data analyst expert. Currently the focus is on management of data resources and data models, associative filtering, hypergraph exploration and aggregations.
Performance
On my 10 years old Intel i3 core machine, TriaClick takes about a minute to load a 42 x 2.8M Physician records TSV flat file (856MB) on SSD and the average elapsed time for processing user selections, i.e. filters with an exploratory QlikView style, is 3 seconds. Output can be transformed and seen as associations, tuples, and columns with distinct values, frequencies and filtering states. The result set can also be driven to a hypergraph for further exploration.