Abstract: The description relates to executing an inference query relative to a database management system, such as a relational database management system. In one example a trained machine learning model can be stored within the database management system. An inference query can be received that applies the trained machine learning model on data local to the database management system. Analysis can be performed on the inference query and the trained machine learning model to generate a unified intermediate representation of the inference query and the trained model. Cross optimization can be performed on the unified intermediate representation. Based upon the cross-optimization, a first portion of the unified intermediate representation to be executed by a database engine of the database management system can be determined, and, a second portion of the unified intermediate representation to be executed by a machine learning runtime can be determined.

Abstract: A method and system for collecting statistics on database tables. In one embodiment, the method includes identifying a plurality of tables for statistics collection, wherein each table of the plurality of tables are remote tables; examining at least one attribute for each table identified for statistics collection; prioritizing the tables for statistics collection based on the at least one attribute; and collecting statistics on the identified tables in a priority sequence.

Abstract: There is disclosed a data processing system implemented method, a data processing system, and an article of manufacture for directing a data processing system to maintain a database table associated with an initial maintenance scheduling interval. The data processing system implemented method includes selecting a randomizing factor, and selecting a new maintenance scheduling interval for the database table based on the initial maintenance scheduling interval and the selected randomizing factor.

Abstract: A system and method are provided for updating database statistics for use in generating query execution plans. A first query for a database is received, for example by a compiler, and the compiler accesses statistics associated with the first query for optimizing a query execution plan for the first query. The statistics may be accessed from a first cache. If the statistics retrieved from the first cache are determined to be stale, the system obtains updated statistics and may store them in a temporary cache. The stale statistics in the first cache are marked as invalid. Existing queries using the first cache may continue to use the first cache, while new queries may be provided with updated statistics.

Abstract: A method, system and computer program product for a database management system to be operatively coupled to a data processing system. The data processing system includes memory for storing a data table. The method, system and computer program product for determining frequency of updating a histogram of the data table. The method includes determining sampled histograms for sample portions of the data table from time to time, computing a difference between at least two of the determined sampled histograms, and determining a frequency with which to update the histogram of the data table depending on the amount of the computed difference.

Abstract: The present invention relates to collecting statistics automatically for data in a database. There is provided a method for automated statistics collection comprising determining a likelihood that statistics for data have changed; and collecting statistics for data in response to the likelihood. Indicators of the likelihood that statistics have changed may be useful to trigger automated statistics collection. Tables having statistics that change significantly may be collected more often than statistics of tables that are stable. A preferred model is provided to facilitate the collection of statistics that are more relevant: a table is scheduled for collection in accordance with observed patterns of table activity; a table is considered for collection if it meets a threshold level of activity; and a table is sampled to predict whether the statistics to be collected have changed. When collecting statistics, throttling and lock contention can minimize impact on a database user's response experience.

Abstract: A system and method are provided for updating database statistics for use in generating query execution plans. A first query for a database is received, for example by a compiler, and the compiler accesses statistics associated with the first query for optimizing a query execution plan for the first query. The statistics may be accessed from a first cache. If the statistics retrieved from the first cache are determined to be stale, the system obtains updated statistics and may store them in a temporary cache. The stale statistics in the first cache are marked as invalid. Existing queries using the first cache may continue to use the first cache, while new queries may be provided with updated statistics.

Abstract: A method and system for collecting statistics on database tables. In one embodiment, the method includes identifying a plurality of tables for statistics collection, wherein each table of the plurality of tables are remote tables; examining at least one attribute for each table identified for statistics collection; prioritizing the tables for statistics collection based on the at least one attribute; and collecting statistics on the identified tables in a priority sequence.

Abstract: There is disclosed a data processing system implemented method, a data processing system, and an article of manufacture for directing a data processing system to maintain a database table associated with an initial maintenance scheduling interval. The data processing system implemented method includes selecting a randomizing factor, and selecting a new maintenance scheduling interval for the database table based on the initial maintenance scheduling interval and the selected randomizing factor.

Abstract: There is disclosed a data processing system implemented method, a data processing system, and an article of manufacture for directing a data processing system to maintain a database table associated with an initial maintenance scheduling interval. The data processing system implemented method includes selecting a randomizing factor, and selecting a new maintenance scheduling interval for the database table based on the initial maintenance scheduling interval and the selected randomizing factor.

Abstract: A method, system and computer program product for a database management system to be operatively coupled to a data processing system. The data processing system includes memory for storing a data table. The method, system and computer program product for determining frequency of updating a histogram of the data table. The method includes determining sampled histograms for sample portions of the data table from time to time, computing a difference between at least two of the determined sampled histograms, and determining a frequency with which to update the histogram of the data table depending on the amount of the computed difference.

Abstract: A method, system and computer program product for a database management system to be operatively coupled to a data processing system. The data processing system includes memory for storing a data table. The method, system and computer program product for determining frequency of updating a histogram of the data table. The method includes determining sampled histograms for sample portions of the data table from time to time, computing a difference between at least two of the determined sampled histograms, and determining a frequency with which to update the histogram of the data table depending on the amount of the computed difference.

Abstract: The present invention provides a method, system and program for optimizing compression of a workload processed by a database management system. In an embodiment of the present invention a method of optimizing the compression of database workloads is provided. Initially, an estimate of a cost of execution for each query according to a defined metric such as execution time or memory consumption is determined. A sub-set of queries is then selected from the workload in order of the most costly to least costly relative to the defined metric for compression according to either a predetermined compression threshold percentage or a threshold percentage derived from an allotted workload execution time. Compression is then performed on the selected sub-set of queries (i.e. those that will benefit the most from the compression) to achieve a net beneficial trade-off between the cost of workload compression and the cost of workload execution.

Abstract: There is disclosed a data processing system implemented method, a data processing system, and an article of manufacture for directing a data processing system to maintain a database table associated with an initial maintenance scheduling interval. The data processing system implemented method includes selecting a randomizing factor, and selecting a new maintenance scheduling interval for the database table based on the initial maintenance scheduling interval and the selected randomizing factor.

Abstract: The present invention relates to collecting statistics automatically for data in a database. There is provided a method for automated statistics collection comprising determining a likelihood that statistics for data have changed; and collecting statistics for data in response to the likelihood. Indicators of the likelihood that statistics have changed may be useful to trigger automated statistics collection. Tables having statistics that change significantly may be collected more often than statistics of tables that are stable. A preferred model is provided to facilitate the collection of statistics that are more relevant: a table is scheduled for collection in accordance with observed patterns of table activity; a table is considered for collection if it meets a threshold level of activity; and a table is sampled to predict whether the statistics to be collected have changed. When collecting statistics, throttling and lock contention can minimize impact on a database user's response experience.

Abstract: A method, system and computer program product for a database management system to be operatively coupled to a data processing system. The data processing system includes memory for storing a data table. The method, system and computer program product for determining frequency of updating a histogram of the data table. The method includes determining sampled histograms for sample portions of the data table from time to time, computing a difference between at least two of the determined sampled histograms, and determining a frequency with which to update the histogram of the data table depending on the amount of the computed difference.

Abstract: The present invention provides a method, system and program for optimizing compression of a workload processed by a database management system. In an embodiment of the present invention a method of optimizing the compression of database workloads is provided. Initially, an estimate of a cost of execution for each query according to a defined metric such as execution time or memory consumption is determined. A sub-set of queries is then selected from the workload in order of the most costly to least costly relative to the defined metric for compression according to either a predetermined compression threshold percentage or a threshold percentage derived from an allotted workload execution time. Compression is then performed on the selected sub-set of queries (i.e. those that will benefit the most from the compression) to achieve a net beneficial trade-off between the cost of workload compression and the cost of workload execution.