Deleting a continuous query from a data stream management system continuing to operate on other queries

• US 7,739,265 B2
• Filed: 10/18/2007
• Issued: 06/15/2010
• Est. Priority Date: 10/18/2007
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Srinivasan, Anand, Jain, Namit, Mishra, Shailendra Kumar
• Primary Examiner(s): Le; Debbie M
• Application Number:
  US11/874,850
• Publication Number:
  US 20090106215A1
• Time in Patent Office: 971 Days
• Field of Search: 707 2- 7, 707/101
• US Class Current: 707/713
• CPC Class Codes: G06F 16/24568 Data stream processing; Continuous queries

Abstract

A first continuous query is deleted from several continuous queries that are being executed, by performing different actions that depend on how resources are shared therebetween. As an example, a second continuous query is deleted if a view is referenced by the first continuous query, and if the view is built on the second continuous query, and if the view is not referenced by any query other than the first continuous query. Such deletion is followed by compiling the second continuous query, followed by updating the view to refer to a newly-compiled version of the second continuous query. As another example, if an operator in the first continuous query is not used by any other query, this operator is de-scheduled from execution, and any exclusively-owned resources are released. Also, data to be read by the operator and data generated by the operator, if present in shared resources, are deleted therefrom.

Claims

1. A computer-implemented method of processing streams of data using continuous queries, the method comprising:

receiving an instruction to delete a first continuous query;

deleting a second continuous query on which is built a view referenced by the first continuous query, if said view is not referenced by any continuous query other than the first continuous query;

compiling by a computer the second continuous query, in response to said deleting; and

updating said view to refer to a newly-compiled version of the second continuous query, in response to said compiling;

if a predetermined condition is met;

de-scheduling execution of an operator used by the first continuous query;

freeing any resource owned by the operator; and

deleting input data still to be processed by the operator if a queue, holding the input data, is shared with another operator, and alternatively marking the queue for deletion if the queue is not shared.

2. The method of claim 1 further comprising:

recursively performing said deleting, said compiling and said updating with another continuous query on which is built another view referenced by the second continuous query, if said another view is referenced by none other than said second continuous query and said first continuous query.

3. The method of claim 1 further comprising:

returning if an output of the first continuous query is used by a view.

4. The method of claim 1 further comprising:

deleting output data generated by the first continuous query, from a store if an owner of the store is shared by other queries and alternatively marking said store for deletion if the store is not shared.

5. A computer-readable storage medium encoded with instructions to process streams of data using continuous queries, the instructions comprising:

instructions to receive an identity of a first continuous query to be deleted;

instructions to check if an output of the first continuous query is used by any view among a predetermined plurality of views, thereby to obtain a first result;

instructions to check if the first continuous query receives input from any view among the predetermined plurality, thereby to obtain a second result;

instructions responsive to each of the first result and second result being false, to further check if an operator used by the first continuous query is also used by any other continuous query, thereby to obtain a third result; and

instructions responsive to the third result being true, to remove any reference to the first continuous query from the operator;

instructions responsive to the third result being false, to de-schedule execution of the operator, to drop any output queue of the operator, and to drop any store of the operator;

instructions to delete input data still to be processed by the operator if a first queue, holding the input data, is shared with another operator and alternatively to mark the first queue for deletion if the first queue is not shared; and

instructions to delete output data generated by the first continuous query, from a store if an owner of the store is shared by other queries and alternatively to mark said store for deletion if the store is not shared.

6. The computer-readable storage medium of claim 5 further comprising:

instructions responsive to the first result being true, to return without changing execution of the first continuous query.

7. The computer-readable storage medium of claim 5 further comprising:

instructions responsive to the second result being true, to check if the view from which the first continuous query receives input is also referenced by any other continuous query not to be deleted, to obtain a fourth result;

instructions responsive to the fourth result being false to delete a second continuous query used in said view;

instructions to compile the second continuous query again, after said deleting; and

instructions to update said view to refer to a compiled version of the second continuous query resulting from said compiling.

8. The computer-readable storage medium of claim 7 further comprising:

instructions responsive to the fourth result being true, to repeat execution of said instructions to check to obtain the second result.

9. A data stream management system comprising a computer and a memory, said data stream management system comprising:

means for receiving an instruction to delete a first continuous query;

means, coupled to the means for receiving, for returning without changing execution of the first continuous query if an output of the first continuous query is used by any view among a predetermined plurality of views;

means, coupled to the means for receiving, for deleting a second continuous query used in a view referenced by the first continuous query if the first continuous query receives input from said view, compiling the second continuous query after deletion, and updating said view to refer to a compiled version of the second continuous query resulting from compilation; and

means, coupled to the means for receiving, for de-scheduling execution of an operator in the first continuous query, dropping any output queue owned by the operator, dropping any store owned by the operator, deleting input data still to be processed by the operator if a first queue holding the input data is shared with another operator and alternatively marking the first queue for deletion if the first queue is not shared, and deleting output data generated by the first continuous query from a store if an owner of the store is shared by other queries and alternatively marking said store for deletion if the store is not shared.

10. The data stream management system of claim 9 further comprising:

means for recursively invoking said means for deleting, compiling and updating with another continuous query on which is built another view referenced by the second continuous query, if said another view is referenced by none other than said second continuous query and said first continuous query.