A Cartesian product is returned when join conditions are not specified in a PROC SQL join. In a Cartesian product, each row from the first table is combined with every row from the second table.

This PROC SQL query is an inner join. It combines the rows from the first table that match rows from the second table, based on the matching criteria specified in the WHERE clause. Columns are not overlaid, so all columns from the referenced tables (including any columns with duplicate names) are displayed. Any unmatched rows from either table are not displayed.

This PROC SQL query is a right outer join, which retrieves all rows that match across tables, based on the join conditions in the ON clause, plus nonmatching rows from the right (second) table.

There are two valid formats for writing a PROC SQL inner join. The PROC SQL query shown at the top of this question uses the first inner join format, which does not use a keyword to indicate the type of join. The alternate format is similar to an outer join and uses the keyword INNER JOIN.

This PROC SQL query is a left outer join, which retrieves all rows that match across tables (based on the join conditions in the ON clause), plus nonmatching rows from the left (first) table. No columns are overlaid, so all columns from both tables are displayed.

Inner joins combine the rows from the first table that match rows from the second table, based on one or more join conditions in the WHERE clause. The columns being matched must have the same data type, but they are not required to have the same name. For joins, the tables being joined can have different numbers of columns, and the rows do not need to be sorted.

Unlike a table, an in-line view exists only during query execution. Because it is temporary, an in-line view can be referenced only in the query in which it is defined.

In order to generate the same output as the DATA step and PRINT steps, the PROC SQL full outer join must use the COALESCE function with the duplicate columns specified as arguments.

A maximum of 256 tables can be combined in a single inner join. If the join involves views (either in-line views or PROC SQL views), it is the number of tables that underlie the views, not the number of views, that counts towards the limit of 256.

The use of summary functions does not require the use of table aliases. All of the other statements about table aliases that are shown here are true.