A processing loop is a group of statements which are executed repeatedly until a stated condition has been satisfied, or as long as a certain condition prevails.
This document covers the following topics:
Processing loops can be subdivided into database loops and non-database loops:
Database processing loops
are those created automatically by Natural to process data selected
from a database as a result of a READ,
FIND or
HISTOGRAM statement.
These statements are described in the section Database Access.
Non-database processing loops
are initiated by the statements REPEAT,
FOR,
CALL FILE,
CALL LOOP,
SORT and
READ WORK FILE.
More than one processing loop may be active at the same time. Loops may be embedded or nested within other loops which remain active (open).
A processing loop must be explicitly closed with a corresponding
END-... statement (for example, END-REPEAT,
END-FOR).
The SORT
statement, which invokes the sort program of the operating system, closes all
active processing loops and initiates a new processing loop.
The following topics are covered below:
With the statements READ,
FIND or
HISTOGRAM, you have
three ways of limiting the number of repetitions of the processing loops
initiated with these statements:
using the session parameter LT,
using a LIMIT statement,
or using a limit
notation in a
READ/FIND/HISTOGRAM statement
itself.
At session level, you can specify the session parameter
LT, which
limits the number of records which may be read in a database processing
loop.
GLOBALS LT=100
This limit applies to all READ,
FIND and
HISTOGRAM statements
in the entire session.
In a program, you can use the LIMIT statement to limit the
number of records which may be read in a database processing loop.
LIMIT 100
The LIMIT statement applies to the remainder of the
program unless it is overridden by another LIMIT statement or
limit notation.
With a READ,
FIND or
HISTOGRAM statement
itself, you can specify the number of records to be read in parentheses
immediately after the statement name.
READ (10) VIEWXYZ BY NAME
This limit notation overrides any other limit in effect, but applies only to the statement in which it is specified.
If the limit set with the LT parameter is smaller
than a limit specified with a LIMIT statement or a limit
notation, the LT limit has priority over any of these
other limits.
Non-database processing loops begin and end based on logical condition criteria or some other specified limiting condition.
The REPEAT
statement is discussed here as representative of a non-database loop
statement.
With the REPEAT statement, you specify one or more
statements which are to be executed repeatedly. Moreover, you can specify a
logical condition, so that the statements are only executed either until or as
long as that condition is met. For this purpose you use an
UNTIL or
WHILE
clause.
If you specify the logical condition
in an UNTIL clause, the REPEAT loop will
continue until the logical condition is met;
in a WHILE clause, the REPEAT loop will
continue as long as the logical condition remains true.
If you specify no logical condition, the REPEAT
loop must be exited with one of the following statements:
ESCAPE
terminates the execution of the processing loop and continues processing
outside the loop (see
below).
STOP stops the
execution of the entire Natural application.
TERMINATE
stops the execution of the Natural application and also ends the Natural
session.
** Example 'REPEAX01': REPEAT
************************************************************************
DEFINE DATA LOCAL
1 MYVIEW VIEW OF EMPLOYEES
2 NAME
2 SALARY (1:1)
*
1 #PAY1 (N8)
END-DEFINE
*
READ (5) MYVIEW BY NAME WHERE SALARY (1) = 30000 THRU 39999
MOVE SALARY (1) TO #PAY1
/*
REPEAT WHILE #PAY1 LT 40000
MULTIPLY #PAY1 BY 1.1
DISPLAY NAME (IS=ON) SALARY (1)(IS=ON) #PAY1
END-REPEAT
/*
SKIP 1
END-READ
END
Output of Program REPEAX01:
Page 1 14-01-14 14:15:54
NAME ANNUAL #PAY1
SALARY
-------------------- ---------- ---------
ADKINSON 34500 37950
41745
33500 36850
40535
36000 39600
43560
AFANASSIEV 37000 40700
ALEXANDER 34500 37950
41745
The ESCAPE
statement is used to terminate the execution of a processing loop based on a
logical condition.
You can place an ESCAPE statement within loops in
conditional IF statement
groups, in break processing statement groups (AT END OF DATA,
AT END OF PAGE,
AT BREAK), or as a
stand-alone statement implementing the basic logical conditions of a
non-database loop.
The ESCAPE statement offers the options TOP
and BOTTOM, which determine where processing is to continue after
the processing loop has been left via the ESCAPE statement:
ESCAPE
TOP is used to continue processing at the top of the
processing loop.
ESCAPE
BOTTOM is used to continue processing with the first
statement following the processing loop.
You can specify several ESCAPE statements within the same
processing loop.
For further details and examples of the ESCAPE statement, see the
Statements documentation.
A database statement can be placed within a database processing loop initiated by another database statement. When database loop-initiating statements are embedded in this way, a "hierarchy" of loops is created, each of which is processed for each record which meets the selection criteria.
Multiple levels of loops can be embedded. For example, non-database loops can be nested one inside the other. Database loops can be nested inside non-database loops. Database and non-database loops can be nested within conditional statement groups.
The following program illustrates a hierarchy of two loops, with one
FIND loop nested or
embedded within another FIND loop.
** Example 'FINDX06': FIND (two FIND statements nested)
************************************************************************
DEFINE DATA LOCAL
1 EMPLOY-VIEW VIEW OF EMPLOYEES
2 CITY
2 NAME
2 PERSONNEL-ID
1 VEH-VIEW VIEW OF VEHICLES
2 MAKE
2 PERSONNEL-ID
END-DEFINE
*
FND1. FIND EMPLOY-VIEW WITH CITY = 'NEW YORK' OR = 'BEVERLEY HILLS'
FIND (1) VEH-VIEW WITH PERSONNEL-ID = PERSONNEL-ID (FND1.)
DISPLAY NOTITLE NAME CITY MAKE
END-FIND
END-FIND
END
The above program selects data from multiple files. The outer
FIND loop selects from the EMPLOYEES file all persons
who live in New York or Beverley Hills. For each record selected in the outer
loop, the inner FIND loop is entered, selecting the car data of
those persons from the VEHICLES file.
Output of Program FINDX06:
NAME CITY MAKE -------------------- -------------------- -------------------- RUBIN NEW YORK FORD OLLE BEVERLEY HILLS GENERAL MOTORS WALLACE NEW YORK MAZDA JONES BEVERLEY HILLS FORD SPEISER BEVERLEY HILLS GENERAL MOTORS
Statement reference notation is used for the following purposes:
Referring to previous statements in a program in order to specify processing over a particular range of data;
Overriding Natural's default referencing;
Documenting.
Any Natural statement which causes a processing loop to be initiated and/or causes data elements in a database to be accessed can be referenced, for example:
When multiple processing loops are used in a program, reference notation is used to uniquely identify the particular database field to be processed by referring back to the statement that originally accessed that field in the database.
If a field can be referenced in such a way, this is indicated in the Referencing Permitted column of the Operand Definition Table in the corresponding statement description (in the Statements documentation). See also User-Defined Variables, Referencing of Database Fields Using (r) Notation.
In addition, reference notation can be specified in some statements. For example:
Without reference notation, an AT START OF DATA, AT
END OF DATA or AT BREAK statement will be related to the
outermost active READ, FIND,
HISTOGRAM, SORT or READ WORK FILE loop.
With reference notation, you can relate it to another active processing
loop.
If reference notation is specified with an
ESCAPE
BOTTOM statement, processing will continue with the first
statement following the processing loop identified by the reference
notation.
Statement reference notation may be specified in the form of a statement reference label or a source-code line number.
Statement reference label
A statement reference label consists of several characters, the last
of which must be a period (.). The period serves to identify the entry as a
label.
A statement that is to be referenced is marked with a label by placing the label at the beginning of the line that contains the statement. For example:
0030 ... 0040 READ1. READ VIEWXYZ BY NAME 0050 ...
In the statement that references the marked statement, the label is placed in parentheses at the location indicated in the statement's syntax diagram (as described in the Statements documentation). For example:
AT BREAK (READ1.) OF NAME
Source-code line number
If source-code line numbers are used for referencing, they must be
specified as 4-digit numbers (leading zeros must not be omitted) and in
parentheses. For example:
AT BREAK (0040) OF NAME
In a statement where the label/line number relates a particular field to a previous statement, the label/line number is placed in parentheses after the field name. For example:
DISPLAY NAME (READ1.) JOB-TITLE (READ1.) MAKE MODEL
Line numbers and labels can be used interchangeably.
See also User-Defined Variables, Referencing of Database Fields Using (r) Notation.
The following program uses source code line numbers (4-digit numbers in parentheses) for referencing.
In this particular example, the line numbers refer to the statements that would be referenced in any case by default.
0010 ** Example 'LABELX01': Labels for READ and FIND loops (line numbers) 0020 ************************************************************************ 0030 DEFINE DATA LOCAL 0040 1 MYVIEW1 VIEW OF EMPLOYEES 0050 2 NAME 0060 2 FIRST-NAME 0070 2 PERSONNEL-ID 0080 1 MYVIEW2 VIEW OF VEHICLES 0090 2 PERSONNEL-ID 0100 2 MAKE 0110 END-DEFINE 0120 * 0130 LIMIT 15 0140 READ MYVIEW1 BY NAME STARTING FROM 'JONES' 0150 FIND MYVIEW2 WITH PERSONNEL-ID = PERSONNEL-ID (0140) 0160 IF NO RECORDS FOUND 0170 MOVE '***NO CAR***' TO MAKE 0180 END-NOREC 0190 DISPLAY NOTITLE NAME (0140) (IS=ON) 0200 FIRST-NAME (0140) (IS=ON) 0210 MAKE (0150) 0220 END-FIND /* (0150) 0230 END-READ /* (0140) 0240 END
The following example illustrates the use of statement reference labels.
It is identical to the previous program, except that labels are used for referencing instead of line numbers.
** Example 'LABELX02': Labels for READ and FIND loops (user labels)
************************************************************************
DEFINE DATA LOCAL
1 MYVIEW1 VIEW OF EMPLOYEES
2 NAME
2 FIRST-NAME
2 PERSONNEL-ID
1 MYVIEW2 VIEW OF VEHICLES
2 PERSONNEL-ID
2 MAKE
END-DEFINE
*
LIMIT 15
RD. READ MYVIEW1 BY NAME STARTING FROM 'JONES'
FD. FIND MYVIEW2 WITH PERSONNEL-ID = PERSONNEL-ID (RD.)
IF NO RECORDS FOUND
MOVE '***NO CAR***' TO MAKE
END-NOREC
DISPLAY NOTITLE NAME (RD.) (IS=ON)
FIRST-NAME (RD.) (IS=ON)
MAKE (FD.)
END-FIND /* (FD.)
END-READ /* (RD.)
END
Both programs produce the following output:
NAME FIRST-NAME MAKE
-------------------- -------------------- --------------------
JONES VIRGINIA CHRYSLER
MARSHA CHRYSLER
CHRYSLER
ROBERT GENERAL MOTORS
LILLY FORD
MG
EDWARD GENERAL MOTORS
LAUREL GENERAL MOTORS
KEVIN DATSUN
GREGORY FORD
JOPER MANFRED ***NO CAR***
JOUSSELIN DANIEL RENAULT
JUBE GABRIEL ***NO CAR***
JUNG ERNST ***NO CAR***
JUNKIN JEREMY ***NO CAR***
KAISER REINER ***NO CAR***
KANT HEIKE ***NO CAR***