Browse Prior Art Database

System and Method for Vertical SQL Search with Stem Parsing

IP.com Disclosure Number: IPCOM000200448D
Publication Date: 2010-Oct-14
Document File: 9 page(s) / 271K

Publishing Venue

The IP.com Prior Art Database

Abstract

We propose a new system and method for vertical SQL searching by stem parsing based on SQL structure and core attributes with dynamic weights. Firstly, A SQL is parsed into a tree according to sql structure whose tree node filled with DDL definition attributes, and then the vertical SQL searching is performed based on the tree node, anchor and node’s attribute dynamic weights according to fuzzy text retrieval technique, finally the SQL structure and attributes compound similarity result-set ranking is returned.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 54% of the total text.

Page 01 of 9

System and Method for Vertical SQL Search with Stem Parsing

SQL is widely used in database query system.

As more and more functional enhancement and older edition maintenance are

made in DB product, and also the improvement of database importance, vertical SQL search would become a popular and valuable technique for both DB enterprise product and end user. One of its requirement is finding DB engine existing fixpack using SQL's core similarity (SQL vertical search) comparison, and also point-to-point customer's service SQL package fixpack advisor is another utilization for vertical SQL search with high business value. In the invention,

we propose a system and method for vertical

SQL searching by stem parsing based on SQL structure and core attributes with dynamic weights.

In this invention,

we propose a new system and method for vertical SQL searching by stem parsing based on SQL structure and

core attributes with dynamic weights. Firstly,

A SQL is parsed into a tree according to sql structure whose tree node filled with DDL

definition attributes, and then the vertical SQL searching is performed based on the tree node, anchor and node's attribute dynamic weights according to fuzzy text retrieval technique, finally the SQL structure and attributes compound similarity result-set ranking is returned.

Claims
1. The system design and component of SQL structure parsing with definition attributes stuffing .
2. The system design and component of dynamic weighted attribute based SQL stem searching .
3. The system design and component of SQL processor (DB engine) fixpack advisor.

Figure 1 gives a general architecture of the proposed system.

1


Page 02 of 9

New Data Corpus (DDL, DML) Rank List of Documents

SQL Stem Extractor

Document Executor:

Merging & Ranking

Stem Matching

Natural Language Learning APPL,DDL,DML Mapping

SQL Stem Indexing

Stem Extractor

Related Web Pages

Figure 1. System Architecture

Online Docs

SQL statement stem could be demonstrated as a tree structure model. Each node in the tree is the key node of the stem,

which

could be the featured target key term from the search perspective. TABLE 1 defines parts of SQL nodes.

TABLE 1

SQL Node Description Example in SQL Statement BQUERY Root node of a SELECT query block SELECT* FROM T1

WHERE T1.C1 = 1

2


Page 03 of 9

INSERT Root node of a INSERT statement INSERTINTO T1 VALUES(…) DELETE Root node of a DELETE statement DELETEFROM T1

WHERE C1 = 1

UPDATE Root node of a UPDATE statement UPDATET1 SET C1 = 2
WHERE C1 = 1

TABLE Table node SELECT * FROM T1 COL Column node C1
VCMP
Value comparing predicate C1 =1
BETWN BETWEEN predicate C1 BETWEEN1 AND 100 BOOL OR, AND predicate connection C1 = 1 ANDC2 = 1

C1 = 1 ORC2 = 1

CALL Call a stored procedure CALL
STOREPROC(VAR1, VAR2, …)

CASE CASE node for CASE … WHEN … ELSE …
END expression

CASEC1

WHEN1 THEN …
WHEN2 THEN …
ELSE… END

WHEN WHEN node for CASE … WHEN … ELSE …
END expression


CAST CAST function CAST(C1 AS VARCHAR(100)) UDF User...