Set this property to define primary keys.

Data Type

string

Default Value

""

Remarks

Google BigQuery does not natively support primary keys, but for certain DML operations or database tools you may need to define them. By default this option is disabled and no tables will have primary keys except for the ones defined in schema files (if you set Location).

Primary keys are defined using a list of rules which match tables and provide a list of key columns. For example, PrimaryKeyIdentifiers="*=key;transactions=tx_date,tx_serial;user_comments=" has three rules separated by semicolons:

1. The first rule *=key means that every table without a more specific rule will have one primary key column called key. Tables that do not have a key column will not have any primary keys.
2. The second rule transactions=tx_date,tx_serial means that the transactions table will have the two primary key columns tx_date and tx_serial. If any of those columns are missing from the table then they will be ignored.
3. The third rule user_comments= means that the user_comments table will have no primary keys. The only use that empty key lists have is in overriding the default rule. If there is no default rule present then the only tables with primary keys would be the ones explicitly listed.

Note that the table names can include just the table, the table and dataset or the table, dataset and project. Both column and table names may be quoted using SQL quotes:

/* Rules with just table names use the connection ProjectId (or DataProjectId) and DatasetId. 
   All these rules refer to the same table with a connection where ProjectId=someProject;DatasetId=someDataset */
someTable=a,b,c
someDataset.someTable=a,b,c
someProject.someDataset.someTable=a,b,c

/* Any table or column name may be quoted */
`someProject`."someDataset".[someTable]=`a`,[b],"c"