Data sampling method and device based on distributed memory database

The invention relates to a data sampling method based on a distributed memory database, which takes the distributed memory database as a filtering container and takes a data filtering rule as a filtering condition. The attributes of the filtering container comprise a distributed cluster server, a data cache size and a data cache strategy. The filtering conditions comprise that a 128-bit HASH value is calculated according to an MD5 algorithm on the basis of a rule, and a data storage memory database organization rule is constructed on the basis of the HASH value, so that data can be rapidly and accurately extracted according to a user-defined rule in front of large data and mass data. According to the method and the system, the PB can set a large number of levels of data sampling effects, and the effect of obtaining required sampling result data within a short time without influencing the service efficiency in the service use process can be met. The situation can be quickly mastered through sampling analysis o