cassandra/doc/modules/cassandra/pages/reference/cql-commands/create-custom-index.adoc

432 lines
11 KiB
Plaintext

= CREATE CUSTOM INDEX
:description: Creates a storage-attached index (SAI).
include::cassandra:partial$sai/support-databases.adoc[]
Creates a Storage-Attached Indexing (SAI) index.
You can create multiple secondary indexes on the same database table, with each SAI index based on any column in the table.
All column date types except the following are supported for SAI indexes:
* `counter`
* non-frozen user-defined type (UDT)
.One exception
****
You cannot define an SAI index based on the partition key when it's comprised of only one column.
If you attempt to create an SAI index in this case, SAI issues an error message.
****
However,you can define an SAI index on one of the columns in a table's composite partition key, i.e., a partition key comprised of multiple columns.
If you need to query based on one of those columns, an SAI index is a helpful option. In fact, you can define an SAI index on each column in a composite partition key, if needed.
Defining one or more SAI indexes based on any column in a database table (with the rules noted above) subsequently gives you the ability to run performant queries that use the indexed column to filter results.
See the xref:cassandra:developing/cql/indexing/sai/sai-overview.adoc[SAI section].
== Syntax
BNF definition:
[source,bnf]
----
include::example$BNF/index_name.bnf[]
----
// tag::syntax[]
----
CREATE CUSTOM INDEX [ IF NOT EXISTS ] [ <index_name> ]
ON [ <keyspace_name>.]<table_name> (<column_name>)
| [ (KEYS(<map_name>)) ]
| [ (VALUES(<map_name>)) ]
| [ (ENTRIES(<map_name>)) ]
USING 'StorageAttachedIndex'
[ WITH OPTIONS = { <option_map> } ] ;
----
// end::syntax[]
.Syntax legend
[%collapsible]
====
include::cassandra:partial$cql-syntax-legend.adoc[]
====
index_name::
Optional identifier for index.
If no name is specified, the default used is `<table_name>\_<column_name>\_idx`.
Enclose in quotes to use special characters or to preserve capitalization.
+
include::cassandra:partial$index-naming.adoc[]
column_name::
The name of the table column on which the SAI index is being defined.
SAI allows only alphanumeric characters and underscores in names.
SAI returns `InvalidRequestException` if you try to define an index on a column name that contains other characters, and does not create the index.
map_name::
// LLP FIX
Used with xref:cassandra:developing/cql/collections/collection-create.adoc[collections], identifier of the `map_name` specified in `CREATE TABLE` ...
`map(<map_name>)`.
The regular column syntax applies for collection types `list` and `set`.
option_map::
Define options in JSON simple format.
+
[cols="100,223"]
|===
| Option | Description
| `case_sensitive`
| Ignore case in matching string values.
Default: `true`.
| `normalize`
| When set to `true`, perform https://unicode.org/faq/normalization.html[Unicode normalization] on indexed strings.
SAI supports Normalization Form C (NFC) Unicode.
When set to `true`, SAI normalizes the different versions of a given Unicode character to a single version, retaining all the marks and symbols in the index.
For example, SAI would change the character Å (U+212B) to Å (U+00C5).
When implementations keep strings in a normalized form, equivalent strings have a unique binary representation.
See https://unicode.org/reports/tr15/[Unicode Standard Annex #15, Unicode Normalization Forms].
Default: `false`.
| `ascii`
| When set to `true`, SAI converts alphabetic, numeric, and symbolic characters that are not in the Basic Latin Unicode block (the first 127 ASCII characters) to the ASCII equivalent, if one exists.
For example, this option changes à to a.
Default: `false`.
| similarity_function
| Vector search relies on computing the similarity or distance between vectors to identify relevant matches.
The similarity function is used to compute the similarity between two vectors.
Valid options are: EUCLIDEAN, DOT_PRODUCT, COSINE
Default: `COSINE`
|===
== Query operators
SAI supports the following query operators for tables with SAI indexes:
include::cassandra:partial$sai/supported-query-operators-list.adoc[]
SAI does not supports the following query operators for tables with SAI indexes:
include::cassandra:partial$sai/notSupportedOperators.adoc[]
== Examples
These examples define SAI indexes for the `cycling.cyclist_semi_pro` table, which is demonstrated in the xref:cassandra:getting-started/sai-quickstart.adoc[SAI quickstart].
[source,language-cql]
----
include::cassandra:example$CQL/sai/cyclist_semi_pro_sai_indices.cql[tag=createQuickStartIndices]
----
For sample queries that find data in `cycling.cyclist_semi_pro` via these sample SAI indexes, see xref:cassandra:getting-started/sai-quickstart.adoc#saiQuickStart__saiQuickStartSubmitQueries[Submit CQL queries].
Also refer xref:cassandra:developing/cql/indexing/sai/sai-query.adoc[Querying with SAI].
[[saiCollectionsExamples]]
=== SAI collection map examples with keys, values, and entries
// LLP FIX
The following examples demonstrate using collection maps of multiple types (`keys`, `values`, `entries`) in SAI indexes.
For related information, see xref:cassandra:developing/cql/collections/collection-create.adoc[Creating collections] and xref:cassandra:developing/cql/collections/map.adoc[Using map type].
Also refer to the SAI collection examples of type xref:#saiCreateCustomIndexCollectionsListAndSetExamples[list and set] in this topic.
First, create the keyspace:
[source,language-cql]
----
CREATE KEYSPACE demo3 WITH REPLICATION =
{'class': 'SimpleStrategy', 'replication_factor': '1'};
----
Next, use the keyspace:
[source,language-cql]
----
USE demo3;
----
Create an `audit` table, with a collection map named `text_map`:
[source,language-cql]
----
CREATE TABLE audit ( id int PRIMARY KEY , text_map map<text, text>);
----
Create multiple SAI indexes on the same `map` column, each using `KEYS`, `VALUES`, and `ENTRIES`.
[source,language-cql]
----
CREATE CUSTOM INDEX ON audit (KEYS(text_map)) USING 'StorageAttachedIndex';
CREATE CUSTOM INDEX ON audit (VALUES(text_map)) USING 'StorageAttachedIndex';
CREATE CUSTOM INDEX ON audit (ENTRIES(text_map)) USING 'StorageAttachedIndex';
----
Insert some data:
[source,language-cql]
----
INSERT INTO audit (id, text_map) values (1, {'Carlos':'Perotti', 'Marcel':'Silva'});
INSERT INTO audit (id, text_map) values (2, {'Giovani':'Pasi', 'Frances':'Giardello'});
INSERT INTO audit (id, text_map) values (3, {'Mark':'Pastore', 'Irene':'Cantona'});
----
Query all data:
[tabs]
====
Query::
+
--
[source, sql]
----
SELECT * FROM audit;
----
--
Result::
+
--
[source,results]
----
id | text_map
----+---------------------------------------------
1 | {'Carlos': 'Perotti', 'Marcel': 'Silva'}
2 | {'Frances': 'Giardello', 'Giovani': 'Pasi'}
3 | {'Irene': 'Cantona', 'Mark': 'Pastore'}
(3 rows)
----
--
====
Query using the SAI index for specific entries in the `map` column:
[tabs]
====
Query::
+
--
[source, sql]
----
SELECT * FROM audit WHERE text_map['Irene'] = 'Cantona' AND text_map['Mark'] = 'Pastore';
----
--
Result::
+
--
[source,results]
----
id | text_map
----+-----------------------------------------
3 | {'Irene': 'Cantona', 'Mark': 'Pastore'}
(1 rows)
----
--
====
Query using the SAI index for specific keys in the `map` column using `CONTAINS KEY`:
[tabs]
====
Query::
+
--
[source, sql]
----
SELECT * FROM audit WHERE text_map CONTAINS KEY 'Giovani';
----
--
Result::
+
--
[source,results]
----
id | text_map
----+---------------------------------------------
2 | {'Frances': 'Giardello', 'Giovani': 'Pasi'}
(1 rows)
----
--
====
Query using the SAI index for specific values in the `map` column with `CONTAINS`:
[tabs]
====
Query::
+
--
[source, sql]
----
SELECT * FROM audit WHERE text_map CONTAINS 'Silva';
----
--
Result::
+
--
[source,results]
----
id | text_map
----+------------------------------------------
1 | {'Carlos': 'Perotti', 'Marcel': 'Silva'}
(1 rows)
----
--
====
// tag::saiCollectionsReferenceNote[]
Remember that in CQL queries using SAI indexes, the `CONTAINS` clauses are supported with, and specific to:
* SAI *collection maps* with `keys`, `values`, and `entries`
* SAI *collections* with `list` and `set` types
// end::saiCollectionsReferenceNote[]
[[saiCreateCustomIndexCollectionsListAndSetExamples]]
=== SAI collection examples with list and set types
These examples demonstrate using collections with the `list` and `set` types in SAI indexes.
For related information, see:
// LLP FIX
* xref:cassandra:developing/cql/collections/collection-create.adoc[Creating collections]
* xref:cassandra:developing/cql/collections/list.adoc[Using list type]
* xref:cassandra:developing/cql/collections/set.adoc[Using set type]
If you have not already, create the keyspace.
[source,language-cql]
----
CREATE KEYSPACE IF NOT EXISTS demo3 WITH REPLICATION =
{'class': 'SimpleStrategy', 'replication_factor': '1'};
----
[source,language-cql]
----
USE demo3;
----
==== Using the list type
Create a `calendar` table with a collection of type `list`.
[source,language-cql]
----
CREATE TABLE calendar (key int PRIMARY KEY, years list<int>);
----
Create an SAI index using the collection's `years` column.
[source,language-cql]
----
CREATE CUSTOM INDEX ON calendar(years) USING 'StorageAttachedIndex';
----
Insert some random `int` list data for `years`, just for demo purposes.
[TIP]
====
Notice the `INSERT` command's square brackets syntax for list values.
[source,language-cql]
----
INSERT INTO calendar (key, years) VALUES (0, [1990,1996]);
INSERT INTO calendar (key, years) VALUES (1, [2000,2010]);
INSERT INTO calendar (key, years) VALUES (2, [2001,1990]);
----
====
Query with `CONTAINS` example:
[tabs]
====
Query::
+
--
[source, sql]
----
SELECT * FROM calendar WHERE years CONTAINS 1990;
----
--
Result::
+
--
[source,results]
----
key | years
-----+--------------
0 | [1990, 1996]
2 | [2001, 1990]
(2 rows)
----
--
====
This example created the `calendar` table with `years list<int>`.
Of course, you could have created the table with `years list<text>`, for example, inserted 'string' values, and queried on the strings.
==== Using the set type
Now create a `calendar2` table with a collection of type `set`.
[source,language-cql]
----
CREATE TABLE calendar2 (key int PRIMARY KEY, years set<int>);
----
Create an SAI index using the collection's `years` column -- this time for the `calendar2` table.
[source,language-cql]
----
CREATE CUSTOM INDEX ON calendar2(years) USING 'StorageAttachedIndex';
----
Insert some random `int` set data for `years`, again just for demo purposes.
[TIP]
--
Notice the `INSERT` command's curly braces syntax for set values.
[source,language-cql]
----
INSERT INTO calendar2 (key, years) VALUES (0, {1990,1996});
INSERT INTO calendar2 (key, years) VALUES (1, {2000,2010});
INSERT INTO calendar2 (key, years) VALUES (2, {2001,1990,2020});
----
--
Query with `CONTAINS` example from the list:
[tabs]
====
Query::
+
--
[source, sql]
----
SELECT * FROM calendar2 WHERE years CONTAINS 1990;
----
--
Result::
+
--
[source,results]
----
key | years
-----+--------------------
0 | {1990, 1996}
2 | {1990, 2001, 2020}
(2 rows)
----
--
====