# Nonclustered Index

![`smart`](NonclusteredIndex05_aa_002_20230922.png)

----------------------------------------------------
We refer to a collection of key-value pairs in the database as a region, which is also the logical unit for data retrieval. Obviously, putting all the data from the entire database into a single region would result in very low reading efficiency. Therefore, when a region reaches a certain size limit, it will undergo splitting. As for the splitting of new Regions, they will be split when they are either written to a size of 96 MB or updated to a size of 144 MB (changed to the value set by the 'region-max-size' TiKV configuration parameter after TiDB 6.1.0). You can refer to the description of Clustered Index Table splitting in the previous section of this course for more details on this concept.