Structured data is a form of data that is normally
represented in terms of records. A record may consist
of multiple field-value pairs. For example, a record of
“Student” may contain a field of “Firstname” with a
value of “สมชาย” along with other field-value pairs
(Figure 1). Structured data is often stored in Relational
Database Management System (RDBMS), which
utilizes SQL as the query language.
Retrieval of text data in DBMS is most effective
when text string stored in each field is short. When a
field contains short text strings, DBMS allows the data
in the field to be indexed based on some efficient data
structure, e.g. B+
-Tree [1]. However, when the stored
text strings are longer than a size limit, the indexing
technique can not be applied and must be disabled.
The size of text string that can be indexed using B+
-
Tree is generally limited by disk block size. Thus,
searching fields that stores long text strings in DBMS
requires combination of sequential search and string
pattern matching technique which results in longer
processing time.