Creates a secondary index on one or more columns of a table.
A secondary index is created on the column(s) in the table as specified. The index is stored in the data dictionary under the given name. The secondary index is used internally by the optimizer to improve the efficiency of a search.
The UNIQUE Option
UNIQUEis specified each index value (i.e. the value of all index columns together) is only allowed once. In this context two null values are considered equal. It is recommended that
UNIQUEis used in the
CREATE TABLEstatement rather than in the
index-nameis specified in its unqualified form, the index will be created in the schema which has the same name as the current ident.
index-nameis specified in its fully qualified form (i.e.
schema-name.index-name) the index will be created in the named schema (in this case, the current ident must be the creator of the specified schema).
The COLLATE Clause
COLLATEclause is specified, the index will be ordered according to the collation specified.
Otherwise, the collation is inherited from the column-definition.
For more information, see the Mimer SQL User's Manual, Collations.
DESCindicate the sort order of the column within the index. If neither is specified, then
ASCis implicit. This makes an index appropriate for queries with a matching
WITHOUT CHECKclauses are used to control whether existing table data should be verified for uniqueness or not when a unique index is created.
WITH CHECKis the default behavior.
WITH CHECKis used but the existing data in the table is not unique, the
CREATE INDEXstatement will fail.
WITHOUT CHECKis used and the existing data in the table is not unique, the
CREATE INDEXstatement will still succeed. (After the index has been created, all new data will be verified for uniqueness.)
Note: For a database with the AUTOUPGRADE attribute enabled, the WITHOUT CHECK option must be used when creating a unique index.
See ALTER DATABASE for more information about AUTOUPGRADE.
Two indexes with the same name cannot belong to the same schema.
An index must belong to the same schema as the table on which it is created.
Secondary indexes may only be created on base tables, not on views.
UNIQUEindexes may only be created on tables in databanks defined with the
WITH/WITHOUT CHECKclause is only valid for unique indexes.
Each column name must identify an existing column of the table. The same column may not be identified more than once.
Mimer SQL can make use of an index in both the forward and backward direction. It is therefore immaterial whether
DESCis specified if all the index columns have the same sorting direction.
Secondary indexes are automatically maintained and are invisible to the user. The index is used automatically when it provides better efficiency.
Any column may be specified as a secondary index.
Table columns that are in the primary key, a unique key or used in a foreign key reference are automatically indexed (in the order in which they are defined in the key). Therefore, explicitly creating an index on these columns will not improve performance.
Consider a table with columns A, B and C of which A and B form the primary key, in that order. An index is automatically created for the column combination A,B. Therefore, there is no advantage in explicitly creating an index on column A or on the column combination A,B. Secondary indexes may, however, be advantageous on column B alone or on combinations such as B,A or A,C.
When you use
ALTER TABLE ... ALTER COLUMNto a create a new collation, the index will remain unchanged.
ExamplesCREATE INDEX cst_date_of_birth ON customers(date_of_birth); CREATE INDEX cst_ename_french ON customers(ename COLLATE french_1);
For more information, see Mimer SQL User's Manual, Creating Secondary Indexes.
Mimer Information Technology AB
Voice: +46 18 780 92 00
Fax: +46 18 780 92 40