Is there any class of algorithms that are the most productive on the DBMS side?

Question

Database

Is there any class of algorithms that are the most productive on the DBMS side?

Answer 1

Just puzzled. The most performant strategy is to query only the indexes, and reduce the "greed" of the query.
The second is data denormalization.
The third use of prepared statement
The fourth logic on the side of the database !!!!!!! very dangerous stuff!

Answer 2

"Threshold/Waterfall" descent from unique value to non-unique value at each step (within one or more queries) or "glossary".

Answer 3

100% yes. For example, you should not do a join on the client side by selecting from one table, and then selecting from the second in a loop (even if the selection goes by index on the second table). It's better to send a single RDBMS query.