Hi Alvin Endratno. Thanks.
And for your question, as per what I have discovered, In SQL databases, there is a process that validates that the inserted data corresponds to the defined schema of the table.
This process sometimes takes time as we validate each data item against the corresponding column. Once we go schemaless with NoSQL, we can save this time. Thus, NoSQL databases may provide better write operations per second as compared to SQL DB.
In the case of read operations, both SQL as well as NoSQL support indexes. Yes, SQL joins are better optimized than MongoDB's aggregate functions but MongoDB can store more complex documents, with all the dependent data together in nested structures. That way you don't have to do joins which optimizes for a certain kind of data retrieval.