For multi-source universes in IDT, there is a SQL92 standard-ish syntax. Any universe objects built on a multi-source universe will have to use this syntax. The Data Federation engine will decompose the SELECT statement and then create one or more SELECT statements to be submitted to the underlying database. Those SQL statements will be created using the syntax of the underlying database.

For example, you create a multi-source universe that has two connections, one to an Oracle 19 database and one to a SQL Server 2016 database. You join tables from the two sources in the data foundation, and then you create a WebI report that uses objects from both sources. WebI will build the query in the multi-source syntax and submit to the data federation engine. The data federation engine will then create an Oracle-standard SQL to submit to the Oracle database, and a SQL Server-standard SQL to submit to the SQL Server database.

With single-source IDT, or with UDT, the universe is built using the syntax of the underlying database -- there is no intermediate layer and no translation. What you see in the SQL view of a WebI report is what's submitted to the database.

Similarly, for FHSQL, you are creating the SQL using the database's native syntax.