Vortex searching

• US 7,774,347 B2
• Filed: 08/30/2007
• Issued: 08/10/2010
• Est. Priority Date: 08/30/2007
• Status: Active Grant

• Current Assignee: Perfect Search Corporation
• Original Assignee: Perfect Search Corporation
• Inventors: Millett, Ronald P., Higgins, John C., Inouye, Dillon K.
• Primary Examiner(s): Vo; Tim T.
• Assistant Examiner(s): MENG, JAU SHYA
• Application Number:
  US11/847,688
• Publication Number:
  US 20090063454A1
• Time in Patent Office: 1,076 Days
• Field of Search: None
• US Class Current: 707/747
• CPC Class Codes: G06F 16/3341 using boolean model

Abstract

Determining intersection points of parameter patterns. Parameter patterns are specified in a query. A method includes identifying a first parameter pattern from the query as occurring less often in the index than one other parameter pattern in the query. The data store is searched until a present location of the data store has been identified as including the first parameter pattern. Then the data store is searched for a location of another parameter pattern. If the present location is identified as including the another parameter pattern, then an indication is provided identifying an intersection. Otherwise, the method includes continuing searching remaining portions of the data store to find a location of the another parameter pattern at a new present location. At least one of the acts of searching above includes eliminating at least a portion of records of the data store from searching without being searched prior to being eliminated.

Claims

1. In a computing environment including a data store including data items, a method for determining one or more intersection points of parameter patterns in data items, the parameter patterns being specified in a query of a data store including records identified by sequentially ordered identifiers, the query including N parameter patterns to attempt to locate one or more records including the N parameter patterns, the method comprising a computing system performing the following:

a) identifying a first parameter pattern from a query, the parameter patterns being specified in a query of a data store including records identified by sequentially ordered identifiers, the query including N parameter patterns to attempt to locate one or more records including the N parameter patterns to find one or more intersection points of parameter patterns in data items, wherein the first parameter pattern occurs less often in an index of records of a data store than at least one other parameter pattern specified in the query;

b) searching the index until a record of the data store has been identified as including the first parameter pattern wherein a present location identifies the record of the data store identified as including the first parameter pattern;

c) then searching the index for another parameter pattern from the query;

d) if the present location is identified as including the another parameter pattern, as a result of searching the index for another parameter pattern, then providing an indication identifying a record at the present location as an intersection;

else

e) continuing searching remaining portions of the index to find the another parameter pattern in the data store at a different location, wherein the present location is then configured to identify the different location of the another parameter pattern in the data store; and

wherein at least one of the acts of searching in acts b), c), and e) comprises eliminating at least a portion of records of the data store from searching without being searched prior to being eliminated using a speed-up data structure that skips comparison of record identifiers by skipping indexed record identifiers correlated to offsets in offset correlation table, wherein the offset correlation table indexes parameter patterns based on record identifier offsets, the offset correlation table including an offset column comprising record identifier offsets and a records column comprising record identifiers corresponding to the offsets, wherein the offsets are a predetermined portion of the corresponding record identifiers to which they correspond, wherein the portions of indexed record identifiers are skipped by skipping comparing indexed record identifiers based on the offset, and wherein eliminating at least a portion of records of the data store from searching without being searched prior to being eliminated is performed in response to an act of determining that eliminating at least a portion of records of the data store from searching without being searched prior to being eliminated would be more efficient, by using less processing power, than actually performing comparisons based on the distance between two record identifiers and the size of the database.

2. The method of claim 1, further comprising repeating acts c-e until a predetermined condition is satisfied.

3. The method of claim 2, wherein the predetermined condition comprises identifying a predetermined number of locations where all N parameter patterns intersect.

4. The method of claim 1, wherein repeating act c-e comprises searching the data store for a present location of a different parameter pattern from the search, wherein the different parameter pattern is different than any other parameter pattern from among the N parameter patterns for which an intersection has been identified.

5. The method of claim 1, wherein the data store has data items organized in a ranked fashion, and wherein searching is performed to cause highest ranked data items to be searched for parameter patters before lower ranked data items.

6. The method of claim 5, wherein data items are ordered by topical relevance.

7. The method of claim 5, wherein data items are ordered by at least one of appropriateness or degree of entity formulating the search.

8. The method of claim 5, wherein data items are ordered by a linguistic profile.

9. The method of claim 1, wherein eliminating at least a portion of records of the data store from searching comprises referencing a pointer directed to a location and wherein at least some locations between a current location and the location pointed to by the pointer are eliminated from searching.

10. The method of claim 9, wherein referencing a pointer comprises referencing an offset correlation table, the offset correlation table comprising portions of index entry reference values correlated to index entry addresses.

11. The method of claim 2, wherein the N parameter patterns specified in the query are absolutely ordered in order from a parameter pattern appearing least frequently in the data store to a parameter pattern appearing most frequently in the data store, and wherein searching the data store for a location of another parameter pattern from the query is performed repeatedly using the absolute ordering to cause lower frequency parameter patterns to be searched before higher frequency parameter patterns.

12. The method of claim 2, wherein the N parameter patterns specified in the query are generally ordered in order from a parameter pattern appearing least frequently in the data store to a parameter pattern appearing most frequently in the data store, and wherein searching the data store for a location of another parameter pattern from the query is performed repeatedly using the general ordering to cause lower frequency parameter patterns to be generally searched before higher frequency parameter patterns.

13. The method of claim 1, wherein at least one parameter pattern is a composite of other parameter patterns in the search.

16. The method of claim 2, wherein when repeating acts c-e, the another parameter pattern in at least some subsequent iterations alternates between the first parameter pattern and a second parameter pattern.

14. In a computing environment including a data store including data items, a computer readable medium storing computer executable instructions that when executed by one or more processors cause the following acts to be performed:

a) identifying a first parameter pattern from a query, the parameter patterns being specified in a query of a data store including records identified by sequentially ordered identifiers, the query including N parameter patterns to attempt to locate one or more records including the N parameter patterns to find one or more intersection points of parameter patterns in data items, wherein the first parameter pattern occurs less often in an index of records of a data store than at least one other parameter pattern specified in the query;

b) searching the index until a record of the data store has been identified as including the first parameter pattern wherein a present location identifies the record of the data store identified as including the first parameter pattern;

c) then searching the index for another parameter pattern from the query;

d) if the present location is identified as including the another parameter pattern, as a result of searching the index for another parameter pattern, then providing an indication identifying a record at the present location as an intersection;

else

e) continuing searching remaining portions of the index to find the another parameter pattern in the data store at a different location, wherein the present location is then configured to identify the different location of the another parameter pattern in the data store; and

wherein at least one of the acts of searching in acts b), c), and e) comprises eliminating at least a portion of records of the data store from searching without being searched prior to being eliminated using a speed-up data structure that skips comparison of record identifiers by skipping indexed record identifiers correlated to offsets in offset correlation table, wherein the offset correlation table indexes parameter patterns based on record identifier offsets, the offset correlation table including an offset column comprising record identifier offsets and a records column comprising record identifiers corresponding to the offsets, wherein the offsets are a predetermined portion of the corresponding record identifiers to which they correspond, wherein the portions of indexed record identifiers are skipped by skipping comparing indexed record identifiers based on the offset, and wherein eliminating at least a portion of records of the data store from searching without being searched prior to being eliminated is performed in response to an act of determining that eliminating at least a portion of records of the data store from searching without being searched prior to being eliminated would be more efficient, by using less processing power, than actually performing comparisons based on the distance between two record identifiers and the size of the database.

15. In a computing environment including a data store including data items, a system for determining one or more intersection points of parameter patterns in data items, the system comprising:

one or more processors;

a computer readable memory coupled to the one or more processors, the computer readable memory storing computer executable instructions that when executed by the one or more processors cause the following acts to be performed;

a) identifying a first parameter pattern from a query, the parameter patterns being specified in a query of a data store including records identified by sequentially ordered identifiers, the query including N parameter patterns to attempt to locate one or more records including the N parameter patterns to find one or more intersection points of parameter patterns in data items, wherein the first parameter pattern occurs less often in an index of records of a data store than at least one other parameter pattern specified in the query;

b) searching the index until a record of the data store has been identified as including the first parameter pattern wherein a present location identifies the record of the data store identified as including the first parameter pattern;

c) then searching the index for another parameter pattern from the query;

d) if the present location is identified as including the another parameter pattern, as a result of searching the index for another parameter pattern, then providing an indication identifying a record at the present location as an intersection;

else

e) continuing searching remaining portions of the index to find the another parameter pattern in the data store at a different location, wherein the present location is then configured to identify the different location of the another parameter pattern in the data store; and

wherein at least one of the acts of searching in acts b), c), and e) comprises eliminating at least a portion of records of the data store from searching without being searched prior to being eliminated using a speed-up data structure that skips comparison of record identifiers by skipping indexed record identifiers correlated to offsets in offset correlation table, wherein the offset correlation table indexes parameter patterns based on record identifier offsets, the offset correlation table including an offset column comprising record identifier offsets and a records column comprising record identifiers corresponding to the offsets, wherein the offsets are a predetermined portion of the corresponding record identifiers to which they correspond, wherein the portions of indexed record identifiers are skipped by skipping comparing indexed record identifiers based on the offset, and wherein eliminating at least a portion of records of the data store from searching without being searched prior to being eliminated is performed in response to an act of determining that eliminating at least a portion of records of the data store from searching without being searched prior to being eliminated would be more efficient, by using less processing power, than actually performing comparisons based on the distance between two record identifiers and the size of the database.