Query pipeline

• US 9,009,139 B2
• Filed: 06/10/2011
• Issued: 04/14/2015
• Est. Priority Date: 06/10/2010
• Status: Expired due to Fees

• Current Assignee: Micro Focus LLC (Open Text Corporation)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Huang, Wei, Zhou, Yizheng, Yu, Bin
• Primary Examiner(s): PHAM, KHANH B
• Application Number:
  US13/699,953
• Publication Number:
  US 20130073573A1
• Time in Patent Office: 1,404 Days
• US Class Current: 707/713
• CPC Class Codes: G06F 16/2453 Query optimisation
• CPC Class Codes: G06F 16/254 Extract, transform and load [ETL] procedures, e.g. ETL data flows in data warehouses

Abstract

A query pipeline is created (514) from a query request. The query pipeline includes multiple query operations including multiple query operators. A first query operator and a second query operator perform first and second query operations on a database (526) and on data outside the database (534). A result from the first query operation in the query pipeline is fed to the second query operation in the query pipeline.

Claims

1. A method of generating a query pipeline in a logging system comprising:

parsing received log data into data chunks;

storing the data chunks in a storage;

generating metadata of the data chunks stored in the storage and storing the metadata in a database;

receiving a query request;

determining query operations based on the query request;

creating, by a processor, a query pipeline including the query operations based on the query request, wherein the query operations include a first query operator and a second query operator to perform a first query operation and a second query operation, wherein a result from the first query operation in the query pipeline is fed to the second query operation in the query pipeline;

identifying a username in the query request to determine a level of access of a user;

modifying the query pipeline, including adding a new operator between the first query operator and the second query operator to limit results from the query pipeline based on the determined level of access of the user;

optimizing, by the processor, the modified query pipeline including at least one of;

rearranging the query operators in the query pipeline, eliminating an operator from the query pipeline, or combining multiple operators in the query pipeline into a single operator; and

processing the query operations in the optimized query pipeline to obtain query results from the metadata in the database and the data chunks in the storage.

2. The method of claim 1, comprising:

dividing the query request into a sequence of query operators including the first query operator and the second query operator, wherein the first query operator relates to executing the first query operation on the database in a secondary storage device, and the second query operator relates to executing the second query operation on data in a primary memory device.

3. The method of claim 2, comprising:

selecting the first query operation to be performed on the database based on amount of consumption of memory when executing the first query operation on the database.

4. The method of claim 2, comprising:

loading a result of the first query operation obtained from the database in the primary memory device; and

performing the second query operation according to the second operator on the loaded result in the primary memory device.

5. The method of claim 4, wherein performing the second query operation comprises:

executing the second query operation in the primary memory device without accessing the secondary storage device.

6. The method of claim 1, wherein parsing a received log data into data chunks comprises:

receiving log data;

separating log data into events;

parsing events into field values;

storing the field values and receipt times in a buffer; and

updating a metadata structure storing metadata for the events.

7. The method of claim 6, wherein parsing a received log data into data chunks further comprises:

generating, based on the metadata, the data chunks from the field values in the buffer; and

sending the data chunks to store in the storage.

8. The method of claim 7, wherein each data chunk includes a unique chunk ID identifying the data chunk, a stripe ID shared among a set of data chunks, and an index local identifier for each field value in the data chunks.

9. The method of claim 7, comprising:

updating a chunk table identifying the data chunks stored in a datafile for the storage, wherein the chunk table is used to retrieve the data chunks from the datafile for the storage if the data chunks are included in results of the query operations.

10. The method of claim 9, comprising:

determining whether all data chunks relevant to the results are indexed in the chunk table;

if all data chunks relevant to the results are not indexed in the chunk table, identifying relevant events from a datafile for the storage, retrieving data chunks for the relevant events from the datafile, and sending the retrieved chunks to the storage.

11. A logging system comprising:

a processor to parse received log data into data chunks, store the data chunks in a storage, generate metadata of the data chunks stored in the storage and store the metadata in a database, receive a query request, create a query pipeline including query operations based on the query request, wherein the query operations include a first query operator and a second query operator to perform a first query operation and a second query operation, identify a username in the query request to determine a level of access of a user;

modify the query pipeline by adding a new operator between the first query operator and the second query operator to limit results from the query pipeline based on the determined level of access of the user, optimize the modified query pipeline including at least one of;

rearranging the query operators in the query pipeline, eliminating an operator from the query pipeline, and combining multiple operators in the query pipeline into a single operator, process the query operations in the optimized query pipeline to obtain query results from the metadata in the database and the data chunks in the storage,

wherein a result from the first query operation in the query pipeline is fed to the second query operation in the query pipeline.

12. The system of claim 11, wherein the processor is to divide the query request into a sequence of query operators including the first query operator and the second operator, wherein the first query operator relates to executing the first query operation on the database-in a secondary storage device, and the second query operator relates to executing the second query operation on data in a primary memory device.

13. A non-transitory computer readable medium storing machine readable instructions that when executed by a processor perform a method comprising:

parsing received log data into data chunks;

storing the data chunks in a storage;

generating metadata of the data chunks stored in the storage and storing the metadata in a relational database;

receiving a query request;

creating a query pipeline including query operations based on the query request, wherein the query operations include a first query operator and a second query operator to perform a first query operation and a second query operation, wherein a result from the first query operation in the query pipeline is fed to the second query operation in the query pipeline;

identifying a username in the query request to determine a level of access of a user;

modifying the query pipeline by adding a new operator between the first query operator and the second query operator to limit results from the query pipeline based on the determined level of access of the user;

optimizing the modified query pipeline, including at least one of;

rearranging the query operators in the query pipeline, eliminating an operator from the query pipeline, and combining multiple operators in the query pipeline into a single operator; and

processing the query operations in the optimized query pipeline to obtain query results from the metadata in the database and the data chunks in the storage.

14. The non-transitory computer readable medium of claim 13, wherein the method comprises:

dividing the query request into a sequence of query operators including the first query operator and the second query operator, wherein the first query operator relates to executing the first query operation on the database in a secondary storage device, and the second query operator relates to executing the second query operation on data in a primary memory device.