1. How does data storage?
The data of the data warehouse is stored in two storage: one is stored in a relational database, and the other is stored in a multi-dimensional manner, that is, the multi-dimensional array.
2, what data is stored?
There are different levels of data in the data warehouse. Generally, data is divided into four levels, early detail-level data, current detail-level data, mild integration level, and high integration level. Different levels are generally referred to as particle size. The greater the particle size, the lower the degree of detail, the higher the integration. The level of division is based on the particle size.
There is also a metadata in the data warehouse, which is data about the data. The data dictionary or system directory in the traditional database is metadata, and the metadata in the data warehouse is in two forms: one is metadata established to conversion from the operational environment to the data warehouse environment, which contains the data source. Various attributes and various attributes at the time of conversion; the other metadata is used to establish a map with the multidimensional model and front-end tool.
3, particle size and segmentation
The particle size is a measure of a high degree of integration of data in the data warehouse. The smaller the particle size, the higher the degree of detail, the more the integration is, the more the number of queries; the larger the particle size, the lower the details, the higher the integration, the less the query.
The segmentation is to disperse the data into the respective physical unit to be able to process the efficiency of data processing to improve data processing, respectively. Data units after data division become slice. The criteria of data segmentation can be determined according to the actual situation, usually segmentation according to the date, region or business field, or the like, or may be divided in accordance with multiple standard combination.
4. Organize the data of the data
Here is a relatively simple case, rotating the integrated document. For example, the data storage unit is divided into days, week, quarter, year. The data records in the day record set; then seven days of data is integrated in the weekly record set, and the data in the weekly record set is stored in the quarterly record concentration in the week, and this method will record the earlier record. The higher the integration of storage, the greater the particle size.