UPDATE table data for complex types (RECORD, Geography, JSON) (Native SQL)

Native SQL — the SQL runs on the BigQuery side (server-side). Prefix with #DirectSQL to use this mode. Runs an UPDATE (or other DML) with complex types such as RECORD (array), Geography, and JSON.

Standard SQL query example

This is the base query accepted by the connector. To execute it in SQL Server, you have to pass it to the Data Gateway via a Linked Server. See how to accomplish this using the examples below.

#DirectSQL 
UPDATE TestDataset.DataTypeTest 
SET ColTime='23:59:59.123456',
 ColGeography=ST_GEOGPOINT(34.150480, -84.233870),
 ColRecord=(1,"AA","Column3 data"),
 ColBigNumeric=1222222222222222222.123456789123456789123456789123456789,
 ColJson= JSON_ARRAY('{"doc":1, "values":[{"id":1},{"id":2}]}') 
WHERE ColInteger=1

Using OPENQUERY in SQL Server

SELECT * FROM OPENQUERY([LS_TO_GOOGLE_BIGQUERY_IN_GATEWAY], '#DirectSQL 
UPDATE TestDataset.DataTypeTest 
SET ColTime=''23:59:59.123456'',
 ColGeography=ST_GEOGPOINT(34.150480, -84.233870),
 ColRecord=(1,"AA","Column3 data"),
 ColBigNumeric=1222222222222222222.123456789123456789123456789123456789,
 ColJson= JSON_ARRAY(''{"doc":1, "values":[{"id":1},{"id":2}]}'') 
WHERE ColInteger=1')

Using EXEC in SQL Server (handling larger SQL text)

The major drawback of OPENQUERY is its inability to incorporate variables within SQL statements. This often leads to the use of cumbersome dynamic SQL (with numerous ticks and escape characters).

Fortunately, starting with SQL 2005 and onwards, you can utilize the EXEC (your_sql) AT [LS_TO_GOOGLE_BIGQUERY_IN_GATEWAY] syntax.

DECLARE @MyQuery NVARCHAR(MAX) = '#DirectSQL 
UPDATE TestDataset.DataTypeTest 
SET ColTime=''23:59:59.123456'',
 ColGeography=ST_GEOGPOINT(34.150480, -84.233870),
 ColRecord=(1,"AA","Column3 data"),
 ColBigNumeric=1222222222222222222.123456789123456789123456789123456789,
 ColJson= JSON_ARRAY(''{"doc":1, "values":[{"id":1},{"id":2}]}'') 
WHERE ColInteger=1'
EXEC (@MyQuery) AT [LS_TO_GOOGLE_BIGQUERY_IN_GATEWAY]