- 一维索引
primary: index on unique field, may be sorted on A
clustering: index on non-unique field, file sorted on A
secondary: file not sorted on A
dense: every tuple is referenced by an entry in the index file
sparse: only some tuples are referenced by index file entries
single-level: tuples are accessed directly from the index file
multi-level: may need to access several index pages to reach tuple
一般情况下,主键索引都是dense的,index file 是 sorted。
-
index
Data file sorted: primary index
Data file sorted but one index for each key values
clustering index
Data file unsorted: secondary index -
B + tree
类似 2 3 4 tree
但是可以有重复node -
n 维索引
-
bit-map index
-
MA hashing
choice vector
多列构成近似一个唯一的hash值
cv是可以配置的
notice: range query 在 hash 情况下 不起优化作用
所有内容都是由规定好的cv生成的对应的hash