Search index limits - Tablestore - Alibaba Cloud Documentation Center

This topic describes the limits on search indexes.

Index limits

The following table describes the limits on the number of rows, total size of data, number of fields, read and write rate, synchronization latency, and Nested fields in a search index, and the limits on the fields in the data table for which the search index is created.

Item	Limit	Description
Number of search indexes	None	We recommend that you create only one search index that contains all fields that you want to retrieve.
Number of rows	100 billion	If your business requirements are not met due to the limit, submit a ticket.
Total size	100 TB	If your business requirements are not met due to the limit, submit a ticket.
Number of fields in a search index	1000	None
Read or write rate	50,000 rows/s	The first time data is written or when a large volume of data is written in a short period of time, Tablestore balances loads within a few minutes. The maximum rate of indexing for Text fields is limited to 10,000 rows/s because the process consumes a large number of CPU resources for tokenization. If your business requirements are not met due to the limit, submit a ticket.
Synchronization latency	3 seconds	In most cases, the synchronization latency is within 3 seconds. If your business requirements are not met due to the limit, submit a ticket. Up to 1 minute is required to initialize a new index.
Array length	256	The maximum number of elements in an array.
Nested field	The name of a Nested field must be unique in a search index.	Up to five levels can be nested in a Nested field. Up to 256 child rows are supported for a Nested field. Up to 100 subfields are supported for a Nested field.
Vector field	The name of a Vector field must be unique in a search index.	A search index can contain up to one Vector field. If your business requirements are not met due to the limit, submit a ticket. Up to 2,048 dimensions are supported for a Vector field.
Total length of values in all primary key columns in the data table for which a search index is created	1,000 bytes	The total length of values in all primary key columns in each row can be up to 1,000 bytes.
Length of the value in a primary key column of the String type in the data table for which a search index is created	1,000 bytes	To index a primary key column of the String type, the column value cannot exceed 1,000 bytes in length.
Length of the value in an attribute column of the String type when you want to index the column as Keyword	4 KB	None
Length of the value in an attribute column of the String type when you want to index the column as Text	2 MB	In most cases, the limit is the same as the length limit on an attribute column in a data table. When you perform fuzzy query based on fuzzy tokenization for a Text field and the value of the Text field exceeds 1,024 characters in length, Tablestore truncates and discards the excess characters, and tokenizes only the first 1,024 characters.

Query limits

Search operation

The following table describes the limits when you call the Search operation to query and analyze data.

Category	Item	Limit	Description
General limits	offset+limit	100000	The default value is 10 and the maximum value is 100 for the limit parameter. If your business requirements are not met due to the limit, submit a ticket. When you call the Search operation to query data of a specific column, you can set the limit parameter to up to 1000 if the column is contained in the search index. To increase the number of returned rows, configure the next_token parameter.
	timeout	10 seconds	None
	CU	None	None
	QPS	100,000	The upper limit for lightweight transaction processing is 100,000 queries per second (QPS). If your business requirements are not met due to the limit, submit a ticket.
	Number of query methods that are specified in a Search call	1,024	Query performance is compromised when complex nested queries are specified in a Search operation. We recommend that you simplify the queries.
Terms query	Number of terms	1,024	You can specify up to 1,024 terms in a terms query.
Wildcard query	Length of a query string	32 characters	A query string in a wildcard query can be up to 32 characters in length.
Prefix query	Length of a query string	1,000 character	A query string in a prefix query can be up to 1,000 characters in length.
Full-text search	Number of highlighted text fragments	500	The total number of highlighted text fragments that can be returned when the conditions specified by the inner_hits and highlight parameters are met.
Full-text search	Number of fields for which you can enable the highlight feature	10	You can enable the highlight feature for up to 10 fields in a query.
k-nearest neighbors (KNN) vector query	Top K query results that have the highest similarity to the vector that you want to query in a KNN vector query	1000	Top 1,000 query results that have the highest similarity to the vector that you want to query can be returned in a KNN vector query.
Aggregation and GroupBy	Number of Aggregations at the same level	5	The number of Aggregations is recalculated each time you add a new Aggregation to SubGroupBy.
	Number of GroupBys at the same level	5	The number of GroupBys is recalculated each time you add a new GroupBy to SubGroupBy.
	Number of nested GroupBys	3	The root GroupBy is counted as a nested level.
	Number of Filters in GroupByFilter	10	None
	Number of groups returned by GroupByField	2,000	None
	Number of internal sources of GroupByComposite	32	The maximum number of columns supported by a multi-column field group.
	Number of Ranges in GroupByRange	100	None
	Number of Ranges in GroupByGeodistance	10	None

ParallelScan operation

The following table describes the limits when you call the ParallelScan operation to query data and export data in parallel.

Category	Item	Limit	Description
General limits	offset+limit	Not supported	When you use parallel scan, you cannot configure the offset and limit parameters. The results that are returned are displayed in chronological order. The maximum value of the limit parameter is 2000.
	CU	None	None
	QPS	None	None
Parallel scan	Maximum number of parallel scan tasks in a ParallelScan call	Equal to the value of the MaxParallel parameter	You can call the ComputeSplits operation to obtain the value of the MaxParallel parameter.
	Valid time of the task (aliveTime)	10 minutes	None
	Maximum number of parallel scan tasks	10	Parallel scan tasks that have the same session ID and the same ScanQuery value are considered one task. For more information, see Parallel scan.