• Journal of KIISE:Databases
• /
• v.30 no.2
• /
• pp.168-175
• /
• 2003

Until now, substring selectivities have been estimated by two steps. First step is to build up a count-suffix tree, which has statistical information about substrings, and second step is to estimate substring selectivity using it. However, it's actually impossible to build up a count-suffix tree from biological sequences because their lengths are too long. So, this paper proposes a novel data structure, count q-gram tree, consisting of fixed length substrings. The Count q-gram tree retains the exact counts of all substrings whose lengths are equal to or less than q and this tree is generated in 0(N) time and in site not subject to total length of all sequences, N. This paper also presents an estimation technique, k-MO. k-MO can choose overlapping length of splitted substrings from a query string, and this choice will affect accuracy of selectivity and query processing time. Experiments show k-MO can estimate very accurately.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2006.11a
• /
• pp.357-360
• /
• 2006

범위 술어에 대한 문자열 선택도 추정은 해당 문자열 범위를 숫자 표현으로 변환 해야 하는 어려움이 있다. 하지만 문자열을 숫자 표현으로 변환할 경우 각각의 바이트에 대한 모든 경우의 수를 모두 고려해야 한다. 따라서 변환 시 문자열 뒷부분에 대한 정보를 고려할 수 없는 문제가 발생한다. 최근 연구되고 있는 부분 문자열에 대한 선택도 추정 방식을 적용할 경우 통계정보와 추정에 대한 연산이 증가되는 단점이 있다. 따라서 이는 범위 술어에 대한 추정만을 위해 사용하기에는 적합하지 않다. 따라서 이 논문에서는 B+ Tree 인덱스의 제한적인 통계정보만을 가지고 범위 술어에 대한 문자열 선택도를 추정하는 방법으로 알티베이스에 구현하였다.

• Journal of KIISE:Databases
• /
• v.34 no.2
• /
• pp.141-152
• /
• 2007

Selectivity estimation has been a crucial component in query optimization in relational databases. While extensive researches have been done on this topic for the predicates of numerical data, only little work has been done for substring predicates. We propose novel suffix tree transform algorithms for this problem. Unlike previous approaches where a full suffix tree is pruned and then an estimation algorithm is employed, we transform a suffix tree into a suffix graph systematically. In our approach, nodes with similar counts are merged while structural information in the original suffix tree is preserved in a controlled manner. We present both an error-bound algorithm and a space-bound algorithm. Experimental results with real life data sets show that our algorithms have lower average relative error than that of the previous works as well as good error distribution characteristics.