摘要:
Multi-stream analytics is disclosed. An example method of multi-stream analytics with a query engine includes punctuating unbounded streaming data into data chunks, each of the data chunks representing a bounded data set in the unbounded streaming data. The method also includes processing one of the data chunks. The method also includes rewinding a query instance for processing another of the data chunks.
摘要:
A method is provided for optimizing computer processes executing on a graphics processing unit (GPU) and a central processing unit (CPU). Process data is subdivided into sequentially processed data and parallel processed data. The parallel processed data is subdivided into a plurality of data blocks assigned to a plurality of processing cores of the GPU. The data blocks on the GPU are processed with other data blocks in parallel on the plurality of processing cores. Sequentially processed data is processed on the CPU. Result data processed on the CPU is returned.
摘要:
Data records containing user feedback regarding at least one offering are received. From among candidate attributes in the received data records, a subset of attributes that relate to user sentiment regarding the at least one offering is selected, where the selecting is according to selection criteria including frequency of occurrence of the candidate attributes, relative amounts of negative and positive feedback associated with the candidate attributes, and time density of feedback associated with the candidate attributes. A visualization of at least one sentiment characteristic of the selected subset of attributes that relate to user sentiment is presented for display.
摘要:
A data stream is stored in a database. An event pipe stores a sliding window of events from the data stream. A query is responded to by utilizing the sliding window of events stored in the event pipe.
摘要:
A system and method for processing data are described. The method quantizes a portion of data processed at a first node within a distributed database processing system to generate one or more quantization tuples, aggregates the tuples at the node to create a plurality of aggregated tuples, and transmits the aggregated tuples to a second node within the distributed database processing system.
摘要:
A method and system for processing a data stream are described. The method executes, until the occurrence of a cut condition, a map function from a set of query processing steps to generate map results for a first portion of the data stream, executes a reduce function from the set of query processing steps to generate history-sensitive data from the map results, and rewinds the set of query processing steps, without termination of processing. The history-sensitive data is maintained for a second execution of the map function and reduce function on a second portion of the data stream.
摘要:
Systems and methods of block computation are disclosed. An example of a method includes buffering a plurality of tuples in a build phase. The method also includes, after buffering a last of the plurality of tuples, applying a computation to all of the buffered tuples in a compute phase. The method also includes returning results of the computation one tuple at a time to a query processing pipeline in a stream-out phase.
摘要:
A system and method for processing data are described. The method quantizes a portion of data processed at a first node within a distributed database processing system to generate one or more quantization tuples, aggregates the tuples at the node to create a plurality of aggregated tuples, and transmits the aggregated tuples to a second node within the distributed database processing system.
摘要:
A method is provided for generating a user-defined function (UDF) on a database engine of a database server. A user-function is sent to the database engine. The user-function includes a relational declaration with application and query logic. When the database engine receives the user-function, the database engine generates a function shell. The UDF processes the relational declaration of the user-defined function. The UDF includes the user-function and the function shell. The function shell returns a final data result of the user-defined function through the shell from the database engine.
摘要:
Methods, systems, and computer program products are provided for performing structured parallel data intensive computations. A SQL-FCF framework that defines a SQL construct based interface for performing the computations is provided. The computations within an application are coded as table valued functions (TVFs). Data for the application is partitioned into data partitions to co-locate and load balance the data partitions and corresponding computations. The TVFs are invoked and applied to corresponding ones of the data partitions in a structured order defined by function control forms (FCF).