Methods of Encoding a Combining Integer Lists in a Computer System, and Computer Software Product for Implementing Such Methods
a computer system and integer technology, applied in the field of methods of handling integer lists, can solve the problems of not computationally efficient, the dimension of the bitmap vector becomes, and the memory space required to store the lists in that form becomes a problem
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Benefits of technology
Problems solved by technology
Method used
Image
Examples
example 1
[0116]As an illustration, consider the following query: find the client name and accident date for all car accidents that incurred damages higher than 900, and group the results according to the client name. The query may be processed as follows. First, all the flat file row-ID lists identified in the accident amount thesaurus entries relating to amounts higher than 900 (the five last rows of FIG. 10G) are merged, which yields the list {0, 1, 3, 5, 6, 7} (or the bitmap vector 110101110000 obtained by a bitwise Boolean OR). Then the intersection of that list with the row-ID list identified in the policy type thesaurus entry relating to the value “car” (the second row of FIG. 10D) is determined. The result list {0, 3, 5, 6} (or bitmap vector 100101100000 obtained by a bitwise Boolean AND) specifies the data graphs that satisfy the query criteria. Finally, the entries of the client name thesaurus (FIG. 10A) are read sequentially and when there is a non-empty intersection between the re...
example 2
[0118]To further illustrate the outstanding performance of the VDG scheme, let us consider the query which consists in identifying any client who has had a car accident before the beginning of the civil year of his or her 35th birthday. In a typical conventional system, all the records of the accident data table of FIG. 3 have to be read to obtain the date attribute and policy link values. For each accident record, the policy data table is read at the row-ID found in the policy link column to obtain the policy type attribute and client link values and then, if the policy type is “car”, another access to the client data table is necessary at the row-ID found in the client link column to obtain the birth year attribute value. The latter value is compared with the date attribute value previously obtained in the accident table to determine whether the criteria of the query are fulfilled.
[0119]If the data tables are sorted beforehand on the basis of the attributes referred to in the quer...
example 3
[0284]An example of such a tree is shown in FIG. 37 in the illustrative case of a query which consists in finding all data graphs relating to accidents undergone by client André or client Max and having a damage amount AA such that 500≦AA≦5000. That tree has three leaves, indicated by broken lines, corresponding to the BETWEEN clauses defined in the query: [André, André] and [Max, Max] for the client name attribute and [500, 5000] for the accident amount attribute. The tree also has two nodes, one for the OR operation between the two CN criteria, and one at the root for the AND operation with the AA criterion.
[0285]The tree decomposition is not unique. The one having the minimum number of nodes is preferably selected.
[0286]The next stage 191 is a tree expansion made by analyzing and splitting the BETWEEN clauses relating to attributes having macroword thesauruses. This is done from the tree obtained in step 190, with reference to the sorted thesaurus word and macroword index files a...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More