NoSQL data model (4) Summary of nosql aggregation, NoSQL
Background in the first three articles, we have introduced three databases that belong to the aggregation model in NoSQL: Key-value type, document type, and column family type. The following analyzes the common points and differences of the three aggregate data models to better understand their respective characteristics. Similarities
The common feature of the three data models for aggregation is that they all use the concept of aggregation, and aggregation has an index key that can be used to find its content. When running on a cluster, aggregation is a key step, because the database must ensure that the aggregated data is stored on the same node. The minimum data unit for an aggregation or "Update" operation. For transaction control, the size of an aggregation operation unit is appropriate.
Differences
The key-value data model regards aggregation as an opaque whole, which means that the entire aggregation can only be identified based on the key, rather than simply querying or obtaining a part of the aggregation.
The aggregation of the document model is transparent to data, so you can query or retrieve only a part of the data. However, because the document does not have a mode, when you want to optimize the storage and obtain partial content in the aggregation, the database is not very good at adjusting the document structure.
The column family model divides aggregation into column families so that the database regards it as a data unit in the row aggregation. The structure of such aggregation has some limitations, but the database can use the advantages of this structure to improve its accessibility.