International Business Machines Corporation

Xiao Feng Meng

Ping Liang

Yu He

Xue Bin Cong

Peng Hui Jiang

Sheng Yan Sun

Xu Zhao

Ye Tao

Embodiments analyze an incoming query for a database, train a query model using machine learning based on the analyzing of the incoming query, determine a model result based on the trained query model; extract a similar information query based on the model result, generate a recommendation for query tuning and a visual impact of the recommendation for the query tuning based on the extracting of the similar information query, and output the recommendation for the query turning and the visual impact of the recommendation for the query tuning.

Recommendations for changes in database query performance

Aggregated log index-based log record identification is provided by maintaining log files of a database system on a plurality log storage devices, the log files including log records of changes to database objects, indexing the log files to obtain a collection of log indexes, where the indexing includes, for each log storage device, indexing log records stored on the log storage device to obtain a respective log index, then based on a request, identifying log storage devices that store log records for servicing the request, the log storage devices storing log indexes, of the collection of log indexes, that index an aggregate set of log records on the log storage devices, obtaining the log indexes from the log storage devices, and aggregating the log indexes to provide an aggregated log index that provides an index of the aggregate set of log records on the log storage devices.

Log record identification using aggregated log indexes

A workload is generated to verify a system. The generating includes parsing one or more statements of an input test case to create one or more parsed structures. A comparison is performed of at least one parsed structure of the one or more parsed structures and at least one workload structure of at least one existing workload to identify one or more workload structures as matching the at least one parsed structure. Runtime data relating to at least the one or more workload structures identified as matching is obtained. The workload is created based on at least one matching workload structure of the one or more workload structures identified as matching, the at least one parsed structure of the one or more parsed structures and the runtime data.

Generating workloads for system verification

A self-learning, self-healing database-management system determines that erroneous input, such as an improper SQL query, has caused a critical database condition. The system parses each input statement into a set of tokens, and then translates each set into lines of a failure script. The script is consolidated by cross-referencing each line to infer relationships between input statements. The system then searches historical database logs for previously entered, error-free input similar to the erroneous input. A degree of similarity between the erroneous input and each previously entered input is determined by a computational method that quantifies similarities between the failure script and each script generated from a previously entered input. The system revises the erroneous input to more closely match the most-similar previously entered input and resubmits the corrected input. The results of the resubmission are used to train the system to more accurately correct future input errors.

Growing community of inventors

Beijing, China

Xiao Feng Meng