Index Merge avoids unnecessary full table scans and does not require establishing a large number of composite indexes. In v5.4.0, Index Merge becomes GA. However, you still need to pay attention to the following Invisible indexes Y Y Y Y Y N Composite PRIMARY KEY Y Y Y Y Y Y Unique indexes Y Y Y Y Y Y Clustered index on integer PRIMARY KEY Y Y Y Y Y Y Clustered index on composite or non-integer key Y Y Y Y case above that the column a from the index was used. This can be useful in cases where you have composite indexes. • operator info shows additional details about the access. See Operator info overview

4 5.3 5.2 5.1 5.0 4.0 Composite PRIMARY KEY Y Y Y Y Y Y Y Y Unique indexes Y Y Y Y Y Y Y Y Clustered index on integer PRIMARY KEY Y Y Y Y Y Y Y Y Clustered index on composite or non-integer key Y 275 • Titan plugin • Invisible indexes • Composite PRIMARY KEY • Unique indexes • Clustered indexes on integer PRIMARY KEY • Clustered indexes on composite or non-integer key 4.5.1.1.4 Other supported which represents the ratings of a book by users. The example creates the table and constructs a composite primary key using book_id and user_id, and creates a clustered index on that primary key. CREATE

Y Y Y Invisible indexes Y Y Y N Composite PRIMARY KEY Y Y Y Y Unique indexes Y Y Y Y Clustered index on integer PRIMARY KEY Y Y Y Y Clustered index on composite or non-integer key Y Y Y N 2.3.3 SQL case above that the column a from the index was used. This can be useful in cases where you have composite indexes. • operator info shows additional details about the access. See Operator info overview conditions of one column. For a multi-dimensional composite in- dex, you can use conditions in multiple columns. For example, regarding the composite index (a, b, c): – When a is an equivalent query,

Y Invisible indexes Y Y Y Y N Composite PRIMARY KEY Y Y Y Y Y Unique indexes Y Y Y Y Y Clustered index on integer PRIMARY KEY Y Y Y Y Y Clustered index on composite or non-integer key Y Y Y Y N 1TiDB case above that the column a from the index was used. This can be useful in cases where you have composite indexes. • operator info shows additional details about the access. See Operator info overview conditions of one column. For a multi-dimensional composite in- dex, you can use conditions in multiple columns. For example, regarding the composite index (a, b, c): – When a is an equivalent query,

Y Y Y Invisible indexes Y Y Y N Composite PRIMARY KEY Y Y Y Y Unique indexes Y Y Y Y Clustered index on integer PRIMARY KEY Y Y Y Y Clustered index on composite or non-integer key Y Y Y N 1TiDB incorrectly case above that the column a from the index was used. This can be useful in cases where you have composite indexes. • operator info shows additional details about the access. See Operator info overview conditions of one column. For a multi-dimensional composite in- dex, you can use conditions in multiple columns. For example, regarding the composite index (a, b, c): – When a is an equivalent query,

is not strict #49137 @YangKeao • Fix the issue that data cannot be inserted when a table has a composite index con- sisting of multi-valued indexes and non-binary type strings #49680 @YangKeao • Fix the buckets to improve scan concurrency E E E E E N N N N N N Invisible indexes Y Y Y Y Y Y Y Y Y Y N Composite PRIMARY KEY Y Y Y Y Y Y Y Y Y Y Y CHECK constraints Y Y N N N N N N N N N Unique indexes Y Y Y 2 5.1 5.0 4.0 Clustered index on integer PRIMARY KEY Y Y Y Y Y Y Y Y Y Y Y Clustered index on composite or non-integer key Y Y Y Y Y Y Y Y Y Y N Multi-valued indexes Y Y Y N N N N N N N N Foreign key

Y Y Y Y Y Y Y Y N Composite PRIMARY KEY Y Y Y Y Y Y Y Y Y Unique indexes Y Y Y Y Y Y Y Y Y Clustered index on integer PRIMARY KEY Y Y Y Y Y Y Y Y Y Clustered index on composite or non-integer key Y which represents the ratings of a book by users. The example creates the table and constructs a composite primary key using book_id and user_id, and creates a clustered index on that primary key. CREATE wasting computing resources caused by frequent data sorting operations. 4.7.4.3 Paging batches for composite primary key tables 4.7.4.3.1 Non-clustered index table For non-clustered index tables (also known

buckets to improve scan concurrency E E E E N N N N N N Invisible indexes Y Y Y Y Y Y Y Y Y N Composite PRIMARY KEY Y Y Y Y Y Y Y Y Y Y CHECK constraints Y N N N N N N N N N Unique indexes Y Y Y Y Y Y Y Y Y Y Y Clustered index on integer PRIMARY KEY Y Y Y Y Y Y Y Y Y Y Clustered index on composite or non-integer key Y Y Y Y Y Y Y Y Y N 1TiDB incorrectly treats latin1 as a subset of utf8. See TiDB which represents the ratings of a book by users. The example creates the table and constructs a composite primary key using book_id and user_id, and creates a clustered index on that primary key. CREATE

to improve scan concurrency E E E E E E E E N N N N Invisible indexes Y Y Y Y Y Y Y Y Y Y Y Y Composite PRIMARY KEY Y Y Y Y Y Y Y Y Y Y Y Y CHECK constraints Y Y Y Y Y N N N N N N N Unique indexes Y 5.3 5.2 5.1 Clustered index on integer PRIMARY KEY Y Y Y Y Y Y Y Y Y Y Y Y Clustered index on composite or non-integer key Y Y Y Y Y Y Y Y Y Y Y Y Multi-valued indexes Y Y Y Y Y Y N N N N N N Foreign which represents the ratings of a book by users. The example creates the table and constructs a composite primary key using book_id and user_id, and creates a clustered index on that primary key. CREATE

to improve scan concurrency E E E E E E E E N N N N Invisible indexes Y Y Y Y Y Y Y Y Y Y Y Y Composite PRIMARY KEY Y Y Y Y Y Y Y Y Y Y Y Y CHECK constraints Y Y Y Y Y N N N N N N N Unique indexes Y 5.3 5.2 5.1 Clustered index on integer PRIMARY KEY Y Y Y Y Y Y Y Y Y Y Y Y Clustered index on composite or non-integer key Y Y Y Y Y Y Y Y Y Y Y Y Multi-valued indexes Y Y Y Y Y Y N N N N N N Foreign which represents the ratings of a book by users. The example creates the table and constructs a composite primary key using book_id and user_id, and creates a clustered index on that primary key. CREATE