Gathering index statistics using sampling

• US 9,177,020 B2
• Filed: 03/22/2014
• Issued: 11/03/2015
• Est. Priority Date: 10/19/2012
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Lashley, Scott D., Miao, Bingjie
• Primary Examiner(s): Jalil, Neveen Abel
• Assistant Examiner(s): TORGRIMSON, TYLER J
• Application Number:
  US14/222,619
• Publication Number:
  US 20140207754A1
• Time in Patent Office: 591 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 16/00 Information retrieval; Database structures therefor; File system structures therefor
• CPC Class Codes: G06F 16/2453 Query optimisation

Abstract

An approach is provided in which a sample point system allocates sample point identifiers to a root node included an index tree that includes multiple leaf nodes. The sample point system distributes the sample point identifiers to the root node'"'"'s child nodes, and recursively traverses through the index tree'"'"'s hierarchical index levels and distributes the sample point identifiers from the child nodes to a subset of the index tree'"'"'s leaf nodes. In turn, the sample point system collects sample data from the subset of the plurality of leaf nodes corresponding to the distributed sample point identifiers.

Claims

1. A method comprising:

allocating, by one or more processors, a plurality of sample point identifiers to a root node included in an index tree corresponding to a database, the index tree including a plurality of leaf nodes;

identifying, by at least one of the one or more processors, a number of first nodes included in a plurality of first nodes that are child nodes of the root node;

computing, by at least one of the one or more processors, a distribution average based upon the number of first nodes and a number of sample point identifiers included in the plurality of sample point identifiers;

distributing, by at least one of the one or more processors, the plurality of sample point identifiers to the plurality of first nodes, wherein an amount of the plurality of sample point identifiers distributed to each of the plurality of first nodes does not exceed the distribution average;

recursively traversing, by at least one of the one or more processors, through a plurality of hierarchical index levels included in the index tree and distributing the plurality of sample point identifiers from the plurality of first nodes to a subset of the plurality of leaf nodes;

collecting sample data, by at least one of the one or more processors, from the subset of the plurality of leaf nodes corresponding to the distributed plurality of sample point identifiers; and

generating, by at least one of the one or more processors, a query plan corresponding to the database based upon the collected sample data.

2. The method of claim 1 wherein:

the distribution is uniformly performed across the plurality of first nodes; and

the subset of the plurality of leaf nodes is uniformly distributed across the plurality of leaf nodes included in the index tree.

3. The method of claim 1 further comprising:

computing a collected sample standard deviation of the collected sample data;

determining that the collected sample standard deviation reaches a standard deviation threshold;

in response to determining that the collected sample standard deviation fails to reach a standard deviation threshold, determining whether a maximum number of iterations is reached;

in response to determining that the maximum number of iterations has not been reached;

invoking a subsequent recursive traversal through the plurality of hierarchical index levels and uniformly distributing the plurality of sample point identifiers to a subsequent subset of the plurality of leaf nodes; and

collecting subsequent sample data from the subsequent subset of the plurality of leaf nodes.

4. The method of claim 3 wherein, in response to determining that the maximum number of iterations has been reached, generating a message that indicates the collected sample data fails to meet the standard deviation threshold.

5. The method of claim 1 further comprising:

generating a plurality of random numbers, wherein each of the plurality of random numbers identify one of the plurality of first nodes; and

wherein the distribution of the plurality of sample point identifiers is based upon the plurality of random numbers and the computed distribution average.

6. The method of claim 1 further comprising:

selecting one of the plurality of first nodes;

determining the number of sample point identifiers that were distributed to the selected first node;

identifying a plurality of second nodes that are child nodes of the selected first node;

computing a subsequent distribution average based upon the number of second nodes and the number of sample point identifiers that were distributed to the selected first node; and

re-distributing the sample point identifiers distributed to the selected first node to the plurality of second nodes based upon the computed subsequent distribution average.

7. The method of claim 6 further comprising:

determining that the plurality of second nodes are included in the plurality of leaf nodes; and

assigning the re-distributed sample point identifiers to the plurality of second nodes.

8. The method of claim 6 further comprising:

determining that the plurality of second nodes fail to be included in the plurality of leaf nodes;

in response to determining that the plurality of second nodes fail to be included in the plurality of leaf nodes;

selecting one of the plurality of second nodes;

determining the number of sample point identifiers that were re-distributed to the selected second node;

identifying a plurality of third nodes that are child nodes of the selected second node;

computing a different distribution average based upon the number of third nodes and the number of sample point identifiers that were distributed to the selected second node; and

re-distributing the sample point identifiers distributed to the selected second node to the plurality of third nodes based upon the computed different distribution average.

9. The method of claim 1 wherein the distribution includes utilizing a plurality of memory arrays to store the plurality of sample point identifiers during the recursive traversal, and wherein each one of the plurality of memory arrays corresponds to one of the plurality of hierarchical index levels, each one of the plurality of memory arrays utilized for each index node included in their corresponding hierarchical index level during the recursive traversal.