This preview shows page 1. Sign up to view the full content.
Unformatted text preview: 2.09 log10 125 16 Index overhead data file 200 MB (=2000x105) Index file: 1st level 2nd level 3rd level Total: 800 blocks (=100000/125) 7 blocks (=800/125) 1 blocks (=7/125) 808 blocks = 1.6 MB 17 Problem with Overflows Accessing the overflow block
encumber the search speed. Primary block serve these space for inserting. 18 clustered file Records with the same indexed keys are physically stored in adjacent locations, so as to reduce disk accesses. For textual retrieval: Compute similarity for every text pairs. 19 FIGURE 14.2 A clustering index on the DEPTNUMBER ordering nonkey field of an EMPLOYEE file. 20 FIGURE 14.3 Clustering index with a separate block cluster for each group of records that share the same value for the clustering field. 21 B+-tree Primary index on SSN for employee records Access: O(logmn/b.f. + 1)
m: the order of a B+ tree n: the number of indexing keys 22 B-tree based file structures B-tree B*-tree Virtual B-tree B+-tree Prefix-B-tree 23 The invention of the B-Tree Bayer and McCreight 1972 logkI: retrieval time proportional to logkI, where I is the size of the index and k is the page size.
log641,000,000 4 log21,000,000 20 -Construction to achieve Balance quality. 24 Formal definition of B+-tree of order m - in general, order m>=k, order means the max. branches a nonleaf node can have) A nonleaf page with k descenda...
View Full Document
This note was uploaded on 10/25/2009 for the course EE 2011 taught by Professor Denny during the Spring '09 term at National Tsing Hua University, China.
- Spring '09