2026-06-22 - Paper on transformation search accepted at TaDA@VLDB26

DBGroup member Rubab Zahra Sarfraz got her work on improving the efficiency of transformation search accepted for the VLDB 2026 Workshop on Tabular Data Analysis (TaDA) [1]. In this work, Rubab presents a method that utilizes lightweight data statistics and operator preconditions to prune large parts of the combinatorial search space for transformations. The method is easy to integrate into any transformation search approach and significantly improves performance. As a proof of concept, we demonstrated several orders-of-magnitude speed-up for the AutoJoin system which discovers transformations for joinability of two input columns.

1. 1 C TaDA Speeding Up Transformation Search with Lightweight Statistics
   Rubab Zahra Sarfraz and Boris Glavic
   VLDB 2024 Workshop: Tabular Data Analysis Workshop (TaDA) (2026).
   • To Appear
   • BibTex
   • doi
   • details

   @inproceedings{SG26a,
     author = {Sarfraz, Rubab Zahra and Glavic, Boris},
     title = {Speeding Up Transformation Search with Lightweight Statistics},
     keywords = {Transformation Search},
     booktitle = {VLDB 2024 Workshop: Tabular Data Analysis Workshop (TaDA)},
     projects = {},
     pages = {},
     doi = {},
     year = {2026},
     venueshort = {{TaDA}},
     istoappear = true
   }
   

   
   
   details