The only way to be able to store different types of data not known in advance is to store them as BLOBs which makes it impossible to use any of the advantages of relational databases later on. This means that the whole structure of the data stored in the database needs to be known in order to create a schema which is required in order to store data.
So the effort to setup comes from the fact that to park a car in there, you need to disassemble the car to be able to store it in the garage.
#NOSQL ARANGODB SERIES IRIS UPDATE#
Whenever you need to be able to store new types of data or whenever you need to store additional attributes, you need to update your schema and make sure that existing data are made compatible to the new schema. The relationships between entities have to be modeled. In a relational database, all entities stored in the database must be defined with a schema known in advance. So what are the main shortcomings of relational databases: Oracle, Sybase ASE, MS SQL Server, MySQL…). But the goal is of course not to get rid of SQL but rather to support use cases which are currently not well supported by classical relational database management system (e.g. So this means it’s a database which can be worked with without using SQL but it doesn’t mean that none of them actually supports an SQL syntax.
NoSQL doesn’t necessarily stand for “No SQL” but rather for “Not only SQL”. I’m currently evaluating which NoSQL we could use in a new project and thought I’d document the considered options are and the relevant criteria here.įirst let’s see what a NoSQL database is at all.