A NoSQL explanation

Key value data Model key value database is a simple hash table, mainly used when all database access is through the primary key to operate the case. When discussing the characteristics of each NoSQL database, be aware of its "consistency", "transactional", query characteristics, data structure, and extensibility.
Document Data Model

Column Family Data Model
Same point

The common denominator of the three aggregation-oriented data models is that they all use aggregation, and there is an index key in the aggregation that can find its contents. When running on a cluster, aggregations are the focus, because the database must ensure that the data within the aggregation exists on the same node. Aggregation is the smallest unit of data for an update operation, and for transaction control, the aggregation is the operating unit, which is the appropriate size.

Different points

The key-value data Model regards aggregation as an opaque whole, which means that the entire aggregation can only be detected based on the key, rather than just querying or getting part of it.

The aggregation of the document model is transparent to the data, so that only a subset of the data can be queried or fetched, but because the document has no schema, the database is not well tuned for the document structure when it wants to optimize storage and get some of the aggregations.

The column family model divides aggregations into column families, so that the database treats them as a unit of data within a row aggregation. There is some limitation to the structure of such aggregations, but the database can take advantage of this structure to improve its accessibility.

Figure

A NoSQL explanation