Checklist for Installing a Function - Aster Execution Engine

A number of steps are required to install a function and make it usable for other analysts. In later sections, we will provide complete descriptions of all the steps, but first let us quickly walk through all the steps needed to install a function and make it usable:

1. Create the function, compile it, and package it (typically into a .jar file). The name of the file you're installing must match the name of the function as you coded it. See Build and Package the SQL/MR Function.

   SQL-MapReduce function names are not case-sensitive. Function names will be automatically converted to all lowercase when they are installed.

   SQL-MapReduce function names within a schema must be unique. If you wish to replace an existing SQL-MapReduce function with a new function of the same name, you must follow these steps:

   1. Remove the existing function by issuing \remove in ACT.
   2. Install the new function using \install.
   3. Set permissions on the new function (see SQL-MapReduce Security).
2. Run ACT and log in as the SQL user who will install and manage the function.
3. Use the \install command to install the function, taking care to specify the schema in which the function will be installed. The command has the syntax:

   \install file_pathname schemaname/installed_filename

   If your installation attempt fails, make sure your SQL user account has the INSTALL FILE and CREATE FUNCTION privileges. See SQL-MapReduce Security.

4. Type \dF+ function-name to check which schema the function belongs to.
5. Use the GRANT command to give the EXECUTE privilege to users who will run the function. The syntax is, roughly:

   GRANT EXECUTE
      ON FUNCTION schema-name.function-name
      TO { user-name | group-name | PUBLIC };