Blocks

ABL is a block-structured language. Blocks contain a set of ABL statements that are executed in sequence and are terminated by an END statement. The top level or outermost block of a procedure file is called the main block. ABL code blocks can be nested. Some ABL statements define blocks that begin with a block-header statement that end with a colon (:), and conclude with an END statement.

The following example code shows a block. Note that the FOR EACH statement starts the block.

FOR EACH Customer:
  DISPLAY CustNum Name.
END.

Blocks are typically used to manage transactions. A transaction is a set of changes to a database, or an undoable temp-table, variable, or class data member, which can be rolled back if something unexpected happens. Transactions are scoped to blocks and every block has default error handling built-in. When a transaction block is nested within another transaction block, it is called a subtransaction. When an error occurs, the first step is to undo the transaction or subtransaction. Thus, undoing a transaction reverses changes to database fields, undoable variables, class data members,and temp-table fields. You can further control error handling by defining explicit block error handling and by adding CATCH and FINALLY statements to the end of blocks.

Blocks are also used to create a scope for a database buffer. (You learn more about database buffers in a later topic.)

There are some general rules for determining what constitutes a block:

• ABL statements such as FOR EACH, DO, and REPEAT define the start of a new block.
• Every procedure itself constitutes a block.
• A database trigger (block of ABL code that executes whenever a specific database event occurs) is a block of its own.
• An internal procedure, user-defined function, constructor, method, and destructor define a block.

The complete list of block types is summarized in the following table:

Procedure Blocks and Data Access

Transactions and trigger and procedure blocks