This document summarizes research on implementing search-as-you-type functionality in relational database forms. It motivates this approach by noting limitations of existing search paradigms like SQL and keyword search. Key challenges include enabling fast prefix matching, synchronizing local and global search results, handling errors and misspellings, and improving scalability for large databases. Initial achievements include a prototype called Seaform-DBLP that supports basic prefix search of a single database table, but has limitations around error tolerance, returning all results rather than top-k, and being memory-resident rather than native to a database system. Overall, search-as-you-type in database forms shows promise for balancing usability and functionality, but addressing

1. Database Research Group Search-As-You-Type in Forms: Leveraging the Usability and the Functionalityof Search Paradigm in Relational Databases Hao WuSupervised by Prof. Lizhu ZhouDatabase Research Group, Tsinghua University VLDB PhD Workshop – Sept. 13, Singapore

2. Motivation Problem Statement Challenges Initial Achievements Conclusions

3. Motivation Problem Statement Challenges Initial Achievements Conclusions

4. Motivation Relational databases are widely used. There are many search paradigms: Structured Query Language (SQL) Keyword Search (KS) Query-By-Example (QBE) Different search paradigms are needed by different users. 10/8/2010 4 Hao Wu, DB Group, Tsinghua University

5. Motivation #1: SQL is complex. SELECT* FROMAuthor A, Autor_Paper AP, Paper P WHERE title LIKE'keyword' AND title LIKE'search' AND authors LIKE'g%' AND A.id = AP.aidAND P.id = AP.pid 10/8/2010 5 Hao Wu, DB Group, Tsinghua University

6. Motivation #2: Traditional keyword search is imprecise. Title? Conf. name? Author name? keyword search g 10/8/2010 6 Hao Wu, DB Group, Tsinghua University

7. Motivation #3: Form is awkward. UCI Directory: http://directory.uci.edu/index.php?form_type=advanced_search 10/8/2010 7 Hao Wu, DB Group, Tsinghua University

8. Motivation #4: The "Search" button is not convenient. 10/8/2010 8 Hao Wu, DB Group, Tsinghua University

9. Motivation + Keyword Search + Form-Style Interface + Search-as-you-type Seaform = 10/8/2010 9 Hao Wu, DB Group, Tsinghua University

10. Motivation Problem Statement Challenges Initial Achievements Conclusions

11. Motivation Problem Statement Challenges Initial Achievements Conclusions

12. Problem Statement Data: Single relational table. Several searchable attributes. 10/8/2010 Hao Wu, DB Group, Tsinghua University 12

13. Problem Statement Query: A set of keywords (prefixes) split by fields. A focus indicator. 10/8/2010 Hao Wu, DB Group, Tsinghua University 13 Title: Author: al Focus = Author xml

14. Problem Statement Results: Global results: corresponding tuples. Local results: corresponding attribute values. Aggregations. 10/8/2010 Hao Wu, DB Group, Tsinghua University 14 xml database (albert) xml search (albert) xml security (alice) Title: Author: al albert2 alice1 xml

15. Motivation Problem Statement Challenges Initial Achievements Conclusions

16. Motivation Problem Statement Challenges Initial Achievements Conclusions

17. Challenges: Search-As-You-Type Prefix matching: E.g.al albert, alice, …Trie structure w/ cache. Fast response: Synchronization of local resultsand global results yields heavycomputational cost.On-demand synchronization and dual-list trie. 10/8/2010 Hao Wu, DB Group, Tsinghua University 17

18. Challenges: Error Tolerance Misplacing of keywords: E.g. input "albert"into the Title input box.Automatic query refinement (given a query, how can we modify it to obtain more results?)Large search space; rely on precise estimation and probabilistic model. Fuzzy matching: E.g. input "albrt" instead of "albert".Edit-distance computation on trie structure.Ranking issue of local results: should local results be sorted by edit-distance, or by aggregation values? 10/8/2010 Hao Wu, DB Group, Tsinghua University 18

19. Challenges: Scalability Handle large-scale databases: There are large number of tuples.1) Top-k algorithmPrecise aggregation is impossible in this case.2) Using RDBMS itselfIndex structure should be redesigned for DBMS; performance issues. Handle multiple tables: Data are regularized to several tables.Generalize the single-table local-global computation and reduce on-the-fly joins using pre-joined tables.It is hard to determine which tables are the most necessary to pre-join; extra storage cost. 10/8/2010 Hao Wu, DB Group, Tsinghua University 19

20. Motivation Problem Statement Challenges Initial Achievements Conclusions

21. Motivation Problem Statement Challenges Initial Achievements Conclusions

23. Prefix matching.

25. Non-top-k, i.e. it returns all matching results.

26. Memory-resident.10/8/2010 22 Hao Wu, DB Group, Tsinghua University

27. Demonstrations: Sept. 14, Tuesday 2 14:00 to 15:30 Sept. 15, Wednesday 5 14:00 to 15:30

28. Motivation Problem Statement Challenges Initial Achievements Conclusions

29. Motivation Problem Statement Challenges Initial Achievements Conclusions

30. Conclusions Search-as-you-type with form is a good choice to balance the usability and functionality. There are still many problems to solve: More effective index other than trie + inverted lists. Support error tolerance. Native DBMS support. Top-k algorithms. Pre-join (materialize) tables. ... 10/8/2010 Hao Wu, DB Group, Tsinghua University 26

31. Thanks http://tastier.cs.thu.edu.cn/seaform/ My homepage: http://dbgroup.cs.thu.edu.cn/wuhao/