NOSQL Data Models No Schema Required Most NOSQL

NOSQL Data Models

No Schema Required Most NOSQL systems don't mandate the use of schemas. Instead the records are self-describing, meaning each record contains the information needed to interpret the record. The self-describing format varies, but it is often a variant of XML or JSON. The disadvantage of not having schemas is larger records and the need for the application (not the database) to interpret the data.

Less Powerful Query Language Most NOSQL query languages are less powerful/expressive than SQL. The languages used then to focus on the retrieval of individual records (or simple sets of records). In particular, few NOSQL systems provide the ability to do joins. This mandates that the application/user must perform the join itself, which is less efficient.

Map Reduce Because the data in NOSQL systems tends to be large, it is often useful to have the DMBS perform data operations before returning results. One common model for data processing that NOSQL systems offer is Map. Reduce. You need to provide two functions: ◦ The map function takes a record and returns a output (often an attribute or Boolean value) ◦ The reduce function takes a list of outputs and results a single object to be returned. The Map-Reduce operation applies the map function to every record, and passes the list of outputs to the reduce function, which is then returned to the user. This allows for fast and parallelized operations on the data that is still in place.

Versioning Some NOSQL systems provide versioning. Versioning is the process of keeping track of a records timestamp (creating and/or modification). This allows for historical queries to be done.

Is versioning needed to have backups? 1. Yes 2. No 3. Depends 4. Depends is always right, right?