Fts4

<init>

Fts4(tokenizer: String, tokenizerArgs: Array<String>, contentEntity: KClass<*>, languageId: String, matchInfo: FtsOptions.MatchInfo, notIndexed: Array<String>, prefix: IntArray, order: FtsOptions.Order)

Marks an Entity annotated class as a FTS4 entity. This class will have a mapping SQLite FTS4 table in the database.

FTS3 and FTS4 are SQLite virtual table modules that allows full-text searches to be performed on a set of documents.

An FTS entity table always has a column named rowid that is the equivalent of an INTEGER PRIMARY KEY index. Therefore, an FTS entity can only have a single field annotated with PrimaryKey, it must be named rowid and must be of INTEGER affinity. The field can be optionally omitted in the class but can still be used in queries.

All fields in an FTS entity are of TEXT affinity, except the for the 'rowid' and 'languageid' fields.

Example:

@Entity
  @Fts4
  public class Mail {
    @PrimaryKey
    @ColumnInfo(name = "rowid")
    private final int rowId;
    private final String subject;
    private final String body;
 
    public Mail(int rowId, String subject, String body) {
        this.rowId = rowId;
        this.subject = subject;
        this.body = body;
    }
 
    public String getRowId() {
        return rowId;
    }
    public String getSubject() {
        return subject;
    }
    public void getBody() {
        return body;
    }
  }
  

contentEntity

val contentEntity: KClass<*>

The external content entity who's mapping table will be used as content for the FTS table.

Declaring this value makes the mapping FTS table of this entity operate in "external content" mode. In such mode the FTS table does not store its own content but instead uses the data in the entity mapped table defined in this value. This option allows FTS4 to forego storing the text being indexed which can be used to achieve significant space savings.

In "external mode" the content table and the FTS table need to be synced. Room will create the necessary triggers to keep the tables in sync. Therefore, all write operations should be performed against the content entity table and not the FTS table.

The content sync triggers created by Room will be removed before migrations are executed and are re-created once migrations are complete. This prevents the triggers from interfering with migrations but means that if data needs to be migrated then write operations might need to be done in both the FTS and content tables.

languageId

val languageId: String

The column name to be used as 'languageid'.

Allows the FTS4 extension to use the defined column name to specify the language stored in each row. When this is defined a field of type INTEGER with the same name must exist in the class.

FTS queries are affected by defining this option, see the languageid= option documentation for details.

matchInfo

val matchInfo: FtsOptions.MatchInfo

The FTS version used to store text matching information.

The default value is MatchInfo#FTS4. Disk space consumption can be reduced by setting this option to FTS3, see the matchinfo= option documentation for details.

notIndexed

val notIndexed: Array<String>

The list of column names on the FTS table that won't be indexed.

order

val order: FtsOptions.Order

The preferred 'rowid' order of the FTS table.

The default value is Order#ASC. If many queries are run against the FTS table use ORDER BY row DESC then it may improve performance to set this option to Order#DESC, enabling the FTS module to store its data in a way that optimizes returning results in descending order by rowid.

prefix

val prefix: IntArray

The list of prefix sizes to index.

tokenizer

val tokenizer: String

The tokenizer to be used in the FTS table.

The default value is FtsOptions#TOKENIZER_SIMPLE. Tokenizer arguments can be defined with tokenizerArgs().

If a custom tokenizer is used, the tokenizer and its arguments are not verified at compile time.

tokenizerArgs

val tokenizerArgs: Array<String>

Optional arguments to configure the defined tokenizer.

Tokenizer arguments consist of an argument name, followed by an "=" character, followed by the option value. For example, separators=. defines the dot character as an additional separator when using the FtsOptions#TOKENIZER_UNICODE61 tokenizer.

The available arguments that can be defined depend on the tokenizer defined, see the SQLite tokernizers documentation for details.