Back to Contents

Language Features

Summary:

• Introduction
• Lettercase Insensitivity
• Whitespace Separators
• Quotation Marks
• Statement Terminator
• Character Set
• Comments
• Program Components
• SQL Support
• Identifiers

Introduction

BDL is an English-like programming language designed for creating relational database applications.

The language includes high-level instructions to implement the user interface of the applications, generate reports, and execute SQL statements to communicate with database servers.

Lettercase Insensitivity

BDL is case insensitive, making no distinction between uppercase and lowercase letters, except within quoted strings. Use pairs of double ( " ) or single ( ' ) quotation marks in the code to preserve the lettercase of character literals, filenames, and names of database entities.

You can mix uppercase and lowercase letters in the identifiers that you assign to language entities, but any uppercase letters in BDL identifiers are automatically shifted to lowercase during compilation.

Tips:

1. It is strongly recommended that you define a naming convention for your projects. For example, you can use underscore notation (get_user_name). If you plan to use the Java notation (getUserName), do not forget that BDL is case insensitive (getusername is the same identifier as getUserName).

Whitespace Separators

BDL is free-form, like C or Pascal, and generally ignores TAB characters, LINEFEED characters, comments, and extra blank spaces between statements or statement elements. You can freely use these whitespace characters to enhance the readability of your source code.

Blank (ASCII 32) characters act as delimiters in some contexts. Blank spaces must separate successive keywords or identifiers, but cannot appear within a keyword or identifier. Pairs of double ( " ) or single ( ' ) quotation marks must delimit any character string that contains a blank (ASCII 32) or other whitespace character, such as LINEFEED or RETURN.

Quotation Marks

In BDL, string literals are delimited by single (') or double (") quotation marks:

'Valid character string'
"Another valid character string"

Do not mix double and single quotation marks as delimiters of the same string. For example, the following is not a valid character string:

'Not A valid character string"

In SQL statements, when accessing a non-Informix relational database, such as a DB2 database from IBM, double quotation marks might not be recognized as database object name delimiters. In the SQL language, the standard specifications recommend to use single quotes for string literals and double quotes for database object identifiers like table or column names. 

To include literal quotation marks within a quoted string, precede each literal quotation mark with the backslash (\), or else enclose the string between a pair of the opposite type of quotation marks:

01 MAIN
02   DISPLAY "Type 'Y' if you want to reformat your disk."
03   DISPLAY 'Type "Y" if you want to reformat your disk.'
04   DISPLAY 'Type \'Y\' if you want to reformat your disk.'
05 END MAIN

A string literal can be written on multiple lines. The compiler merges lines by removing the new-line character.

For more details, see String Literals.

Escape Symbols

The compiler treats a backslash ( \ ) as the default escape symbol, and treats the immediately following symbol as a literal, rather than as having special significance. To specify anything that includes a literal backslash, enter double ( \\ ) backslashes wherever a single backslash is required. Similarly, use \\\\ to represent a literal double backslash.

Statement Terminator

BDL requires no statement terminators, but you can use the semicolon ( ; ) as a statement terminator in some case, like in PREPARE and PRINT statements.

01 MAIN
02   DISPLAY "Hello, World"  DISPLAY "Hello, World"
03   DISPLAY "Hello, World"; DISPLAY "Hello, World"
04 END MAIN

Character Set

The language requires the ASCII character set, but also supports characters from the client locale in data values, identifiers, form specifications, and reports.

Comments

A comment is text in the source code to assist human readers, but which BDL ignores. (This meaning of comment is unrelated to the COMMENTS attribute in a form, or to the OPTIONS COMMENT LINE statement, both of which control on-screen text displays to assist users of the application.)

You can indicate comments in any of several ways:

• A comment can begin with the left-brace ( { ) and end with the right-brace ( } ) symbol. These can be on the same line or on different lines.
• The pound ( # ) symbol (sometimes called the "sharp symbol") can begin a comment that terminates at the end of the same line.
• You can use a pair of minus signs ( -- ) to begin a comment that terminates at the end of the current line. (This comment indicator conforms to the ANSI standard for SQL.)

The compiler treats the --# characters as a whitespace character, rather than as the beginning of a comment. This feature provides backward compatibility for source code modules that begin lines containing BDL extensions to Informix 4GL syntax with those symbols.

Warnings:

1. Within a quoted string, 4GL interprets comment indicators as literal characters, rather than as comment indicators.
2. You cannot use braces ( { } ) to nest comments within comments.
3. Comments cannot appear in the SCREEN section of a form specification file.
4. The # symbol cannot indicate comments in an SQL statement block, nor in the text of a prepared statement.
5. You cannot specify consecutive minus signs ( -- ) in arithmetic expressions, because BDL interprets what follows as a comment. Instead, use a blank space or parentheses to separate consecutive arithmetic minus signs.
6. The symbol that immediately follows the -- comment indicator must not be the sharp (#) symbol, unless you intend to compile the same source file with the Informix 4GL product.

Tips:

1. For clarity and to simplify program maintenance, it is recommended that you document your code by including comments in your source files.
2. You can use comment indicators during program development to disable statements without deleting them from your source code modules.

Program Components

BDL programs are built from source code files with the language compiler, form compiler, and message compiler. Source code files can be:

• Form Specification Files (.per)
• Database Schema Files (.sch, .att, .val)
• Message Files (.msg)
• Source String Files (.str)
• Program Sources Files (.4gl)

In Form Specification Files, you define the layout of application screens. See Forms for more details.

The Database Schema Files describe the structure of the database tables. See Database Schema for more details.

The Message Files hold texts that can be loaded at runtime. Each text is identified by a number. See Message Files for more details.

The Localized Strings allow to customize application strings, which are loaded automatically at runtime. Each string is identified by an identifier. See Localized Strings for more details.

In Program Source Files, you define the structure of the program with instruction blocks ( like MAIN, FUNCTION or REPORT ). The program starts from the MAIN block. The instruction blocks contain BDL instructions that are be executed by the runtime system in the order they appear in the code. Program blocks cannot be nested; neither can any program block cannot be divided among more than one source code module.

Some BDL instructions can include other instructions. Such instructions are called compound statements. Every compound statement of BDL supports the END keyword to mark the end of the compound statement construct within the source code module. Most compound statements also support the EXIT statement keywords, to transfer control of execution to the statement that follows the END statement keywords, where statement is the name of the compound statement. By definition, every compound statement can contain at least one statement block, a group of one or more consecutive SQL statements or other BDL statements. In the syntax diagram of a compound statement, a statement block always includes this element.

SQL Support

A limited syntax of SQL is supported directly by the BDL compiler, so you can write common SQL statements like SELECT, INSERT, UPDATE or DELETE directly in your source code:

01 MAIN
02   DEFINE n INTEGER, s CHAR(20)
03   DATABASE stores
04   LET s = "Sansino"
05   SELECT COUNT(*) INTO n FROM customer WHERE custname = s
06   DISPLAY "Rows found: " || n
07 END MAIN

For SQL statements that have a syntax that is not supported directly by the compiler, the language provides SQL statement preparation from strings like other languages.

01 MAIN
02   DEFINE txt CHAR(20)
03   DATABASE stores
04   LET txt = "SET DATE_FORMAT = YMD"
05   PREPARE sh FROM txt
06   EXECUTE sh
07 END MAIN

For more details about SQL statement preparation, see the Dynamic SQL Instructions.

Identifiers

A BDL identifier is a character string that is declared as the name of a program entity. In the default (U.S. English) locale, every 4GL identifier must conform to the following rules:

• It must include at least one character, without any limitation in size.
• Only ASCII letters, digits, and underscore ( _ ) symbols are valid.
• Blanks, hyphens, and other non-alphanumeric characters are not allowed.
• The initial character must be a letter or an underscore.
• Identifiers are not case sensitive, so my_Var and MY_vaR both denote the same identifier.

Within non-English locales, however, BDL identifiers can include non-ASCII characters in identifiers, if those characters are defined in the code set of the locale that CLIENT_LOCALE specifies. In multibyte East Asian locales that support languages whose written form is not alphabet-based, such as Chinese, Japanese, or Korean, an identifier does not need to begin with a letter.