It enhances the TPC-DS benchmark with complex data distribution and challenging yet semantically meaningful query templates

the joins in the TPC-DS benchmark are mostly between a primary key and a foreign key or between a primary key and another primary key.

Needs

1. distinct query instances
2. the distribution of query templates
3. more evaluation metrics

the paper preserves the realistic database schema from TPC-DS, and it enriches the TPC-DS benchmark in a number of ways:

• more skews and correlations to individual columns and multiple columns within a table and across table
• new and semantically meaningful query templates (non primary-key-foreign-key)
• fne-grained data slicing. This also increases the total number of distinct query instances

Thus, the DSB benchmark enhances the data generation of the TPC-DS benchmark with additional skews and correlations

---

• ← Previous Post
• Next Post →