The essential function of indexing is to improve query efficiency. The underlying index is a data structure that can improve query efficiency. Generally, B+ trees are used. The reason for using B+ trees is that it can ensure that data must be found within three times. The leaf node is also a two-way linked list, which supports scope search. Indexes are divided into primary key index and non-primary key index. In primary key index, it is recommended to use self-increment method of primary keys, because each data inserted can be put to the end, thereby reducing unnecessary overhead (such as rotating trees, etc.). There are also ordinary indexes. In ordinary indexes, the value placed by the leaf nodes is often the value of the primary key. If the primary key is not stored but the data is placed, then when the data changes, not only the normal index itself needs to be maintained, but also the primary key index needs to be maintained, which increases unnecessary overhead. Enumeration values, etc. themselves are not suitable for creating indexes, because data can be quickly found by looking up directly on the table. Indexes will affect the efficiency of addition, deletion and modification, so there is no need to add indexes on data that are not often searched.