• Table - Data source for model. When working in Kyligence, we will first need to synchronize table metadata from data source (typically Hive) before creating a model and loading data. Table metadata includes table name, columns, column type, etc.

  • Model - Logical semantic layer. Model is a set of connected tables and their join relationships. It also defines the fact table, dimension tables, dimensions, measures, and indexes. Kyligence finishes data loading to models and precomputation with the index building jobs. The system supports star schema and snowflake schema as the base of multi-dimensional analysis.

  • Index - Aggregate index and table index. Index is built during data loading process and is used to accelerate query executions.

  • Load Data - The action to load data to Kyligence. Data is also precomputed based on the defined index during the loading process. Each data load will generate one segment. This precomputation greatly accelerates the model query.

    • Rebuild Index - For data already loaded to Kyligence, we will need to rebuild the index if the definition changes. With AI-Augmented engine enabled, Kyligence will provide model and index optimization recommendations based on historical queries. The rebuild index task will be triggered after users accept the recommendations.
  • Recommendation - Based on the query history and the characteristics of source data, the system will provide model and index optimization recommendations.

  • AI-Augmented Engine - Powered by machine learning, the AI-augmented engine uses the actual characteristics of business data and user queries to predict business analysis scenarios and provide recommendations on model designs. By accepting these recommendations, users can optimize models quickly.

Copyright © Kyligence Inc. all right reserved,powered by GitbookLast Modified: 2022-06-29 17:20:31

results matching ""

    No results matching ""