Apache Hive : HPL/SQL - SELECT Statement

SELECT statement allows you to run queries.

SELECT TOP n

You can specify a SELECT statement with the TOP clause. PL/SQL automatically converts it to LIMIT clause for Hive.

Example:

SELECT TOP 3 name FROM sales;    -- SELECT name FROM sales LIMIT 3 is executed

Compatibility: Microsoft SQL Server.

SELECT Without FROM Clause

You can specify a SELECT statement without FROM. PL/SQL automatically adds the FROM clause using the table name defined by the hplsql.dual.table option.

By default, HPL/SQL uses default.dual table name. Make sure such table exists in the database and contains exactly one row and one column only.

Example:

SELECT unix_timestamp();    -- SELECT unix_timestamp() FROM default.dual is executed

Compatibility: Microsoft SQL Server, PostgreSQL, MySQL and Netezza.

FROM TABLE (VALUES … ) Clause

You can generate table rows using TABLE (VALUES …) clause (row constructor).

Syntax:

FROM TABLE (VALUES row, ...) AS table_alias (col, ...)

Example 1:

SELECT * FROM TABLE (VALUES 1, 2, 3) AS rn(rownum);

rownum
------
1
2
3

Example 2:

SELECT * FROM (VALUES (1,1), (2,2), (3,3)) AS rn(c1, c2);

c1     c2
---------
1      1
2      2
3      3

Note that HPL/SQL transforms FROM (VALUES …) clause to SELECT UNION ALL subquery before the execution in Hive:

SELECT * FROM
(SELECT 1 AS c1, 1 AS c2 FROM default.dual
 UNION ALL
 SELECT 2 AS c1, 2 AS c2 FROM default.dual
 UNION ALL
 SELECT 3 AS c1, 3 AS c2 FROM default.dual) rn

You can specify the single row, single column dual table using hplsql.dual.table option.

Compatibility: IBM DB2, Microsoft SQL Server

See also:

• SELECT INTO