2results about How to "Improve I/O efficiency" patented technology

The invention provides a data batch migration method which can be applied to the technical field of big data. The method comprises the following steps: creating a thread pool for data batch migration by utilizing a dynamic multi-thread model; performing data slicing on the source data table based on the data volume of the source data table to obtain data batch information, and creating a batch manager with a synchronous lock mechanism; migrating the source data table into a temporary data table in batches by using threads in a thread pool based on the data batch information; the number of threads in the data batch migration process is dynamically adjusted through a dynamic multi-thread model, and the data batch migration process is synchronously controlled through a batch manager; under the condition that the batch manager detects that synchronization is completed, data consistency verification is conducted on the data in the temporary table, and then data batch migration of the source data table is completed. The invention further provides a data batch migration device and equipment, a medium and a program product.

The invention provides a graph construction method and device, a storage medium and computer equipment, and the method comprises the steps: obtaining an original vector graph corresponding to each data vector in a data set; according to the first compression vector of each data vector, constructing a compression vector pool and storing the compression vector pool to a target position; the target position is selected from a memory and a disk according to a preset cooperation strategy; according to the data vectors corresponding to the nodes in the original vector diagram, the node connection relation and the target page node capacity, clustering the nodes to obtain a page node diagram; the page node graph comprises a plurality of page nodes and page connection relations of the page nodes; one page node comprises nodes corresponding to a plurality of data vectors respectively, and different page nodes comprise different nodes; and according to the page connection relationship and the data vector corresponding to each node included in each page node, determining a solid state disk (SSD) page corresponding to each page node.