Index hot loading method and device on basis of advertisement putting

Abstract

The embodiment of the invention provides an index hot loading method and device on the basis of advertisement putting. The index hot loading method on the basis of advertisement putting includes the steps of performing region segmentation on a directory bitmap file to obtain a plurality of index blocks, wherein the directory bitmap file is an advertisement plan index table formed by a plurality ofadvertisement materials and a plurality of advertisement conditions, and each index block includes different advertisement plans; when the directory bitmap file is updated, successively re-constructing a backup index block corresponding to each index block; successively updating the corresponding index block in the directory bitmap file by using each backup index block. By performing region segmentation on the directory bitmap file, the index blocks are obtained so that the memory usage can be greatly reduced each time when a backup index block is reconstructed, after the construction is completed, the corresponding index blocks in the directory bitmap file are replaced by the backup index blocks, the retrieval of other index blocks is not affected in the replacing and updating process, and therefore the retrieval performance is ensured.

Description

technical field [0001] The present invention relates to the technical field of advertisement indexing, in particular to an index hot loading method and device based on advertisement placement. Background technique [0002] In the prior art, the retrieval of basic directional conditions when implementing an advertisement delivery system is usually indexed by means of a directory bitmap file bitmap. Using this type of data structure to construct an index basically meets the requirements of delivery data volume and delivery performance in the field of advertising. However, since the bitmap index will pre-open a piece of memory for storage at the beginning of its establishment, if the index is updated in large quantities, there will be a risk of dirty reads or program crashes in a multi-threaded environment, and it is necessary to add a mutex to ensure data consistency. In this way, index retrieval operations will be affected when updating indexes in large batches. Therefore, t...

AI Technical Summary

Problems solved by technology

[0004] In the first method, the advertising indexing service has multiple nodes, and each node needs to be restarted in turn each time to ensure that the online retrieval service is highly available when updating the index. Difference

[0005] In the second method, the index is rebuilt by creating a backup bitmap. If the amount of data in the advertising plan is large, it will cause huge memory overhead to store two bitmap data at the same time, and the memory redundancy is twice that of the general system.

[0006] In the process of implementing the present invention, the inventor found that there are at least the following problems in the prior art: it is impossible to guarantee that the retrieval index will not be affected when the index is updated, and the memory usage should be controlled as much as possible

Images