This topic describes the data types of column values, API operations, parameters used in the Search operation, query features, usage, and billing of search indexes.

Data types of column values

The value of a field in a search index is the value of the column of the same name from the corresponding table. The data types of the two values must match. The following table describes the matching rules.

Data type of the field value in the search index Data type of the column value in the data table Description
Long Integer 64-bit long integers.
Double Double 64-bit double-precision floating-point numbers.
Boolean Boolean Boolean values.
Keyword String Character strings that cannot be tokenized.
Text String

Character strings or text that can be tokenized. For more information, see Tokenization.

Geo-point String The coordinate information of the location. This parameter value must be in the format of latitude,longitude. Valid values of the latitude: [-90,+90]. Valid values of the longitude: [-180,+180]. Example: 35.8,-45.91.
Nested String The nested type. Example: {"a": 1}, {"a": 3}].
Notice The data types of the values in the search index and the data table must match each other based on the rules described in the preceding table. Otherwise, Tablestore discards the data as dirty data. Make sure that column values of the GEOPOINT and NESTED types comply with the formats described in the preceding table. If the formats do not match, Tablestore discards the data as dirty data. In this case, the data may be available in the data table but unavailable in the search index.

The following table describes the additional attributes of fields in search indexes.

Attribute Type Description
Index Boolean Specifies whether to enable indexing.
  • True indicates that Tablestore creates an inverted index or spatial index for the column.
  • False indicates that Tablestore does not create an index for the column.

    If the column is not indexed, you cannot query data based on the column.

EnableSortAndAgg Boolean Specifies whether to enable the sorting and aggregation features.
  • True indicates that data can be sorted and aggregated by using the column.
  • False indicates that data cannot be sorted or aggregated by using the column.
Store Boolean Specifies whether to store the value of the column in the search index.

True indicates that Tablestore stores the value of the column in the search index. You can query the search index for the value of the column without querying the data table. This improves the query performance.

IsArray Boolean Specifies whether the value of the column is an array.

True indicates that the value of the column is an array. Data written to the column must be a JSON array, such as ["a","b","c"].

You do not need to specify the Array parameter for a nested column because a nested column is an array.

ARRAY type data can be used in all queries because arrays do not affect queries.

isVirtualField Boolean Specifies whether the column is a virtual column.

If you set this parameter to true, the column is a virtual column. If the column is a virtual column, you must specify the source field that is mapped to the column. The source field is from the data table. For more information about virtual columns, see Virtual columns.

The following table describes the combinations of field types and field attributes.
Type Index EnableSortAndAgg Store IsArray
Long Supported Supported Supported Supported
Double Supported Supported Supported Supported
Boolean Supported Supported Supported Supported
keyword Supported Supported Supported Supported
Text Supported Not supported Supported Supported
Geo-point Supported Supported Supported Supported
Nested Applies only to child fields. Applies only to child fields. Applies only to child fields. Nested fields are arrays.

Operations

Operation Description
CreateSearchIndex Creates a search index.
DescribeSearchIndex Queries detailed information of a search index.
ListSearchIndex Queries the list of search indexes.
DeleteSearchIndex Deletes a specified search index.
Search Implements all query features and analysis features such as sorting and aggregation. The results are returned in a specified order.
ParallelScan Implements all query features. You cannot call this operation to sort or aggregate data. Data that matches the query conditions is returned more quickly.

Search

You must specify SearchRequest in a query. The following table describes the parameters that are included in SearchRequest.
Parameter Type Description
offset Integer The position from which the current query starts.
limit Integer The maximum number of items that the current query returns.
getTotalCount Boolean Specifies whether to return the total number of matched rows. By default, this parameter is set to false. A value of true may affect the query performance.
Sort Sort Specifies the field and method for sorting.
Collapse Collapse Specifies the name of the field that you want to collapse in the query result.
query Query The following table describes the query methods that the search index feature supports.
Name Query Description
Match all query MatchAllQuery This query is used to query the total number of rows in a table or retrieve several rows from a data table.
Query by tokenized data MatchQuery This query uses approximate matches to retrieve query results. The keyword you use for a query is tokenized and then is used to match the query results.

The relationship between different tokenized keywords is Or. When the tokenized keywords are partially matched, the row data meets the query conditions.

Match phrase query MatchPhraseQuery This query is similar to match query, but the row data meets the query conditions only when the tokenized words match the row data in the same sequence and position.
Term query TermQuery This query uses full and exact matches to retrieve query results, which is similar to string matching.

If a TEXT column is queried, the row data meets the query conditions when one of the tokenized words can be exactly matched.

Terms query TermsQuery This query is similar to term query, but you can specify multiple words at a time. If one of the words is matched, the row data meets the query conditions.
Prefix query PrefixQuery This query retrieves data that contains the specified prefix.

If a TEXT column is queried, the row data meets the query conditions when one of the tokenized words contains the specified prefix.

Range query RangeQuery This query retrieves data within a specified range.

If a TEXT column is queried, the row data meets the query conditions when one of the tokenized words is within the specified range.

Wildcard query WildcardQuery This query retrieves data that matches a string that contains one or more wildcard characters.

You can use asterisks (*) and question marks (?) as wildcards. An asterisk (*) indicates a sequence of characters.A question mark (?) indicates a single character.

Boolean query BoolQuery You can use BoolQuery to query data based on a combination of filter conditions. This query contains one or more subqueries as filter conditions. Tablestore returns the rows that match the subqueries.

Subquery conditions can be combined by using logical operators such as AND, NOT, and OR.

Nested query NestedQuery You can use a nested query to query the data of nested fields.
Geo-distance query GeoDistanceQuery You can specify a circular geographical area as a filter condition in a query, which consists of a central point and radius. Tablestore returns the rows where the value of a column falls within the geographical circular area.
Geo-bounding box query GeoBoundingBoxQuery You can specify a rectangular geographical area as a filter condition in a query. Tablestore returns the rows where the value of a column falls within the rectangular geographical area.
Geo-polygon query GeoPolygonQuery You can specify a polygonal geographical area as a filter condition in a query. Tablestore returns the rows where the value of a field falls within the polygonal geographical area.
Exists query ExistQuery This query is also called a NULL query or NULL-value query. This query is used for sparse data to determine whether a column in a row exists. For example, you can query the rows in which the value of the address column is not empty.

If a column does not exist in the row or the value of the column is an empty array ("[]"), you can determine that the column does not exist.

Use Tablestore SDKs

You can use the following Tablestore SDKs to implement the search index feature:

Billing

For more information, see Billable items of search indexes.