Framework for supporting regular expression-based pattern matching in data streams

• US 9,305,238 B2
• Filed: 08/26/2009
• Issued: 04/05/2016
• Est. Priority Date: 08/29/2008
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Srinivasan, Anand, Komuravelli, Rakesh, Jain, Namit, Mishra, Shailendra
• Primary Examiner(s): Fleurantin, Jean B
• Assistant Examiner(s): ASPINWALL, EVAN S
• Application Number:
  US12/548,187
• Publication Number:
  US 20100057735A1
• Time in Patent Office: 2,414 Days
• Field of Search: 707/713, 707/758, 707/759, 707/760, 707/761, 707/762, 707/763, 707/765, 707/769, 707/771, 707/772, 707/779, 707/E17.014, 707/E17.017, 707/E17.061, 707/E17.062, 707/E17.063, 707/E17.069, 707/999, 707/3, 707/780, 712/220
• US Class Current: 1/1
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06F 18/20 Analysing

Abstract

Techniques for detecting patterns in one or more data or event streams. A pattern to be detected may be specified using a regular expression. Events received in a data stream are processed during runtime to detect occurrences of the specified pattern in the data stream. In one embodiment, a pattern type or class is determined for the specified pattern and pattern matching is performed using a technique selected based upon the type or class determined for the specified pattern.

Claims

1. A method of detecting a pattern in a data stream comprising a sequence of events, the method comprising:

receiving, at a computerized processing system, a query that specifies a define component that specifies, for each particular symbol in a plurality of symbols in a symbol set, a separate condition that a particular event from the data stream must satisfy in order for the particular symbol to match the particular event;

receiving, at the processing system, a regular expression from the query, the regular expression specifying a pattern of symbols from the symbol set;

selecting, at a central processing unit of the processing system, a particular pattern type, based at least in part on the pattern of symbols specified in the regular expression, wherein selecting the particular pattern type comprises;

selecting the particular pattern type to be a second pattern type and not a first pattern type if any symbols in the regular expression are linked using any operator other than a concatenation operator;

based on the particular pattern type, selecting, at the processing system, a particular technique from among (a) a first technique that concurrently maintains only one binding per state in a finite state automaton constructed based on the regular expression and (b) a second technique that concurrently maintains multiple bindings per state in the finite state automaton; and

using the particular technique for detecting the pattern in the data stream;

wherein selection of the first pattern type as the particular pattern type causes the first technique to be selected as the particular technique;

wherein selection of the second pattern type as the particular pattern type causes the second technique to be selected as the particular technique;

wherein each binding maintained represents at least a partial match of the regular expression by one or more events from the data stream.

2. The method of claim 1 further comprising:

receiving a plurality of events in the data stream, the plurality of events received in sequence;

constructing the finite state automaton for the regular expression; and

detecting a presence of the pattern in the data stream using the automaton by processing each event in the data stream in an order in which the event is received and without re-processing previously received events.

3. The method of claim 1 wherein selecting the particular pattern type comprises:

selecting the particular pattern type to be the second pattern type and not the first pattern type if, for any symbol in the symbol set, a predicate defined for the symbol is dependent on a symbol, in the symbol set, other than the symbol for which the predicate is defined; and

selecting the particular pattern type to be the second pattern type and not the first pattern type if, for any particular symbol in the regular expression having a quantifier that is associated with the particular symbol, the quantifier is not one of (1) a quantifier indicating zero or more occurrences of the particular symbol, (2) a quantifier indicating one or more occurrences of the particular symbol, or (3) a quantifier indicating zero or more occurrences of the particular symbol;

wherein selection of the first pattern type as the particular pattern type causes the first technique to be selected as the particular technique; and

wherein selection of the second pattern type as the particular pattern type causes the second technique to be selected as the particular technique.

4. The method of claim 1, wherein the define component specifies, for at least a particular symbol from the symbol set, a condition that involves a comparison of (a) an attribute of a particular event from the data stream with (b) an attribute of an event preceding the particular event in the data stream.

5. A non-transitory computer-readable medium storing a plurality of instructions for controlling a data processor to detect a pattern in a data stream comprising a sequence of events, the plurality of instructions comprising:

instructions that cause the processor to receive a query that specifies a define component that specifies, for each particular symbol in a plurality of symbols in a symbol set, a separate condition that a particular event from the data stream must satisfy in order for the particular symbol to match the particular event;

instructions that cause the processor to receive a regular expression from the query, the regular expression specifying a pattern of symbols from the symbol set;

instructions that cause the processor to select a particular pattern type, based at least in part on the pattern of symbols specified in the regular expression, wherein the instructions to select the particular pattern type comprise;

instructions that cause the processor to select the particular pattern type to be a second pattern type and not a first pattern type if any symbols in the regular expression are linked using any operator other than a concatenation operator;

instructions that cause the processor to select, based at least in part on the particular pattern type, a particular technique from among (a) a first technique that concurrently maintains only one binding per state in a finite state automaton constructed based on the regular expression and (b) a second technique that concurrently maintains multiple bindings per state in the finite state automaton; and

instructions that cause the processor to use the particular technique for detecting the pattern in the data stream;

wherein the instructions that cause the processor to select the first pattern type as the particular pattern type causes the first technique to be selected as the particular technique;

wherein the instructions that cause the processor to select the second pattern type as the particular pattern type causes the second technique to be selected as the particular technique;

wherein each binding maintained represents at least a partial match of the regular expression by one or more events from the data stream.

6. The non-transitory computer-readable medium of claim 5, wherein the plurality of instructions further comprise:

instructions that cause the processor to receive a plurality of events in the data stream in sequence;

instructions that cause the processor to construct the finite state automaton for the regular expression; and

instructions that cause the processor to detect a presence of the pattern in the data stream using the automaton by processing each event in the data stream in an order in which the event is received and without re-processing previously received events.

7. The non-transitory computer-readable medium of claim 5, wherein the instructions that cause the processor to select the pattern type comprise:

instructions that cause the processor to select the particular pattern type to be the second pattern type and not the first pattern type if, for any symbol in the symbol set, a predicate defined for the symbol is dependent on a symbol, in the symbol set, other than the symbol for which the predicate is defined; and

instructions that cause the processor to select the particular pattern type to be the second pattern type and not the first pattern type if, for any particular symbol in the regular expression having a quantifier that is associated with the particular symbol, the quantifier is not one of (1) a quantifier indicating zero or more occurrences of the particular symbol, (2) a quantifier indicating one or more occurrences of the particular symbol, or (3) a quantifier indicating zero or more occurrences of the particular symbol;

wherein selection of the first pattern type as the particular pattern type causes the first technique to be selected as the particular technique; and

wherein selection of the second pattern type as the particular pattern type causes the second technique to be selected as the particular technique.

8. The non-transitory computer-readable medium of claim 5, wherein the define component specifies, for at least a particular symbol from the symbol set, a condition that involves a comparison of (a) an attribute of a particular event from the data stream with (b) an attribute of an event preceding the particular event in the data stream.

9. A system for detecting a pattern in a data stream, the system comprising:

a memory storing a plurality of instructions; and

a processor coupled to the memory and configured to execute the plurality instructions to;

receive a query that specifies a define component that specifies, for each particular symbol in a plurality of symbols in a symbol set, a separate condition that a particular event from the data stream must satisfy in order for the particular symbol to match the particular event;

receive a regular expression from the query, the regular expression specifying a pattern of symbols from the symbol set;

select a particular pattern type, based at least in part on the pattern of symbols specified in the regular expression, wherein the instructions to select the particular pattern type comprise;

instructions to select the particular pattern type to be a second pattern type and not a first pattern type if any symbols in the regular expression are linked using any operator other than a concatenation operator;

select, based at least in part on the particular pattern type, a particular technique from among (a) a first technique that concurrently maintains only one binding per state in a finite state automaton constructed based on the regular expression and (b) a second technique that concurrently maintains multiple bindings per state in the finite state automaton; and

use the particular technique for detecting the pattern in the data stream;

wherein selection of the first pattern type as the particular pattern type causes the first technique to be selected as the particular technique;

wherein selection of the second pattern type as the particular pattern type causes the second technique to be selected as the particular technique;

wherein each binding maintained represents at least a partial match of the regular expression by one or more events from the data stream.

10. The system of claim 9, wherein the processor is configured to execute the plurality of instructions to:

receive a plurality of events in the data stream in sequence;

construct the finite state automaton for the regular expression; and

detect a presence of the pattern in the data stream using the automaton by processing each event in the data stream in an order in which the event is received and without re-processing previously received events.

11. The system of claim 9, wherein the processor is configured to select the pattern type by being configured to:

select the particular pattern type to be the second pattern type and not the first pattern type if, for any symbol in the symbol set, a predicate defined for the symbol is dependent on a symbol, in the symbol set, other than the symbol for which the predicate is defined; and

select the particular pattern type to be the second pattern type and not the first pattern type if, for any particular symbol in the regular expression having a quantifier that is associated with the particular symbol, the quantifier is not one of (1) a quantifier indicating zero or more occurrences of the particular symbol, (2) a quantifier indicating one or more occurrences of the particular symbol, or (3) a quantifier indicating zero or more occurrences of the particular symbol;

wherein selection of the first pattern type as the particular pattern type causes the first technique to be selected as the particular technique; and

wherein selection of the second pattern type as the particular pattern type causes the second technique to be selected as the particular technique.

12. The system of claim 9, wherein the define component specifies, for at least a particular symbol from the symbol set, a condition that involves a comparison of (a) an attribute of a particular event from the data stream with (b) an attribute of an event preceding the particular event in the data stream.