this is for holding javascript data
H.S.Bhathiya(110059X) edited section_Query_Optimization_and_Execution__.tex
about 8 years ago
Commit id: 37900202eaf6117d87c635c456be230035fc0f78
deletions | additions
diff --git a/section_Query_Optimization_and_Execution__.tex b/section_Query_Optimization_and_Execution__.tex
index 90667e0..988ea4f 100644
--- a/section_Query_Optimization_and_Execution__.tex
+++ b/section_Query_Optimization_and_Execution__.tex
...
Two possible execution plans are deep-left tree (shown in Figure-1 ) and Bushy Tree (shown in Figure-2). Other possible structures include right-deep trees zig-zag trees. Similar to deep-left trees right-deep and zig-zag are linear tree structure. But bushy trees are different and more complicated compared to others.
\subsection{Challenegs in Bushy Tree Implementation}
Moving from left-deep tree to a Bushy tree is a challenge as the number of possible structures in bushy trees are much larger. Left-deep trees have only one structure regardless of the number of tables involved. Therefor the number of possible permutations are N! . But for bushy trees possible number of structures
are is given by
\\
$$ S(N) = 1 if $N \ = \ 1$ $$
\\
$$ $S(N) = \displaystyle \sum_{N=1}^{i} (N)(N-i)$ \ $
if for N > 1$ $$\\
Therefore the number of possible permutations are $S(N)*N!$. Unlike left-deep tree case, estimating the cost for all the possible bushy trees is computationally infeasible for moderately large N (~7). Therefore it is required to come with a heuristic to select set of permutations for cost computations.
\subsection{Heuristics to Select Bushy Trees}