Documents are updated quickly for normalized data and slowly for denormalized data, much as in standard RDBMSes. Reading documents, on the other hand, is faster in denormalized data and slower in normalized data. Denormalized data is more difficult to maintain and takes up more room.
It should be noted that in MongoDB, denormalized data is more commonly anticipated. This is because RDBMSes have built-in support for normalization and enable data to be handled as a separate issue, while NoSQL DBMSes like MongoDB does not.
Instead, normalization necessitates that client applications carefully protect their own integrity. To help with this, audits may be performed to ensure the app data conforms to anticipated patterns of referential integrity.
