Defining Google BigQuery Table Indexes

The following properties are applicable to a Google BigQuery Search Index object.

Tab	Property	Description	Additional Information
	Physical Name	Specifies the name of the search index
	Type	Specifies whether the index type is Search or Vector index
	If Not Exists	Specifies that if the search index already exists, no exception is thrown, and no action happens
	Physical Only	Specifies whether the search index appears in physical models only
	Generate	Specifies whether a DDL statement is generated for the dataset during forward engineering
General	Index Members	Specifies the list of eligible columns to create search or vector index	Available only if All Columns is not selected.
	All Columns	Specifies whether the index is created on all columns of the table	Search Indexes are created only on columns of the types STRING and JSON, or contain a STRING or JSON field in them.
	Analyzer	Specifies controls of the text analyzer and how data is tokenized for searching and indexing
	Data Types	Specifies the data types to use when a search index is created
	Distance Type	Specifies to use a default distance type when performing a vector search	Available when the Type is set to Vector. Valid values are EUCLIDEAN, COSINE, and DOT_PRODUCT.
	Index Type	Specifies the algorithm that is used to build the vector index	Available when the Type is set to Vector.
	Ivf Option	Specifies the algorithm to build the index as an inverted file index	Available when the Type is set to Vector.
	Tree Ah Option	Specifies the algorithm to build the index as a tree-quantization based index	Available when the Type is set to Vector.
	Use Replace Syntax	Specifies whether to use a replace syntax	Available when the Type is set to Vector.
Vector Index	Stored Columns	Specifies the vector indexes for the stored columns	This tab is applicable only when the index Type is set to Vector.
Comments		Specifies comments about the search index

You can create multiple search index per Google BigQuery database table. You can only generate one search index at a time.