InnoDB clustered index is clustered according to the primary key (primary key), each table can only have one clustered index, the table data file itself is an index structure organized by B+Tree, and the data domain of the leaf node stores the complete data record; so, the InnoDB table data file itself is the main index file, that is, you just said that "the B+Tree index and data rows are saved in the same structure". The way the clustered index is used makes it very fast to find and sort according to the range of the primary key (refer to the data structure of the clustered index).

InnoDB secondary indexes are implemented by all secondary indexes refer to the primary key as the data field, so the secondary index search needs to retrieve the secondary index twice to obtain the data record, but it is convenient to change the secondary index (it does not affect the data file itself organized according to the primary index), and because all secondary indexes refer to the primary index, it is not recommended that the primary index be too large.