PROVIDE

PROVIDE Syntax Diagram

Basic form

PROVIDE f1 f2 ... FROM itab1
g1 g2 ... FROM itab2
...
FROM itabi
...
BETWEEN f AND g.

Effect

Retrieves the contents of the specified fields from the internal tables (itab1, itab2, ...) and places them in the table header lines within the required range. Also executes the processing block enclosed by the PROVIDE and ENDPROVIDE statements for each range.

Note

For itab1, itab2 ... only tables with header lines are allowed.


The diagram below illustrates the functionality of the PROVIDE statement for the most simple case where just two tables A and B are to be processed:

IA1 IA2
|-----------| |--------------| table A
: : : :
: IB1 : IB2 : :
: |-----------| |-------------| : table B
: : : : : : : :
: : PROVIDE area : : :
...|----------------------------------------|...
: : : : : : : :
:TI1: TI2 :TI3: : TI4 : TI5 : TI6 :
...|---|-------|---| |-------|-----|-----|...
result ranges

The data structures which form the basis for the table lines must each contain two components which can be interpreted as a range (e.g. start date and end date). In the diagram, the ranges belonging to the entries in table A are marked with IA1 or IA2 , and those in table B with IB1 or IB2. If you split the ranges of both tables into overlapping and non-overlapping ranges and then form the intersection with the PROVIDE area, this results in 6 sub-ranges TI1 to TI6. In these sub-ranges, the values of the tables A and B are constant. The PROVIDE statement makes the contents of the tables A and B available for the 6 sub-ranges, one after the other. It thus acts as a kind of loop where the data of the tables involved can be processed with reference to each range.

Effect

General principle
Each of the specified internal tables has two fields which contain the line-related validity range. You can determine these in the DATA statement with the addition "VALID BETWEEN ... AND ...". If this addition is not used, the first two sub-fields of the table determine these range fields (corresponds to VALID BETWEEN first field AND second field). These fields can be date fields, time fields or number fields. Both these two fields and also f and g should be the same type.

PROVIDE splits the range f to g into sub-ranges so that each of the fields (f1, f2, ...) specified for each table is constant in this range and so that each sub-range is as large as possible (range limits are considered part of the range).
Each time the processing passes through the loop, the current range limits and the specified sub-fields are placed in the header lines of the internal tables. If you want to make all sub-fields available, enter '*' instead of the field list. The unspecified sub-fields are set to their initial value ( CLEAR).
It is a requirement that the ranges within a table are in ascending order and not overlapping. However, there can be gaps between one upper range limit and the next lower range limit.

For each table itab1, itab2 ... , the automatically generated fields itab1_VALID, itab2_VALID, ... indicate (with 'X' or blank ' ') whether a suitable entry was found for the current sub-range.

Example

The entries in the table SE, PR and SH contain time ranges and are filled as follows:

DATA: BEGIN OF SE OCCURS 3, 
		FROM	 TYPE D, 
		TO	 TYPE D, 
		NAME(15) TYPE C, 
		AGE TYPE I, 
	END OF SE, 
 
	BEGIN OF PR OCCURS 4, 
		START	TYPE D, 
		END	TYPE D, 
		PRICE	TYPE I, 
		NAME(10) TYPE C, 
	END OF PR, 
 
	BEGIN OF SH OCCURS 2, 
		CLOSED   TYPE D, 
		STR(20)  TYPE C, 
		OPENED   TYPE D, 
	END OF SH VALID BETWEEN OPENED AND CLOSED, 
 
	BEGIN TYPE D VALUE '19910701', 
	END   TYPE D VALUE '19921001'. 
 
SE-FROM = '19910801'. SE-TO  = '19910930'. 
SE-NAME = 'Shorty'.   SE-AGE = 19. APPEND SE. 
SE-FROM = '19911005'. SE-TO  = '19920315'. 
SE-NAME = 'Snowman'.  SE-AGE = 35. APPEND SE. 
SE-FROM = '19920318'. SE-TO  = '19921231'. 
SE-NAME = 'Tom'.	SE-AGE = 25. APPEND SE. 
 
PR-START = '19910901'. PR-END = '19911130'. 
PR-NAME  = 'Car'.   PR-PRICE = 30000. APPEND PR. 
PR-START = '19911201'. PR-END = '19920315'. 
PR-NAME  = 'Wood'.  PR-PRICE = 10.	APPEND PR. 
PR-START = '19920318'. PR-END = '19920801'. 
PR-NAME  = 'TV'.	PR-PRICE = 1000.  APPEND PR. 
PR-START = '19920802'. PR-END = '19921031'. 
PR-NAME  = 'Medal'. PR-PRICE = 5000.  APPEND PR. 
 
SH-CLOSED = '19920315'. SH-STR = 'Gold Avenue'. 
SH-OPENED = '19910801'. APPEND SH. 
SH-CLOSED = '19921031'. SH-STR = 'Wall Street'. 
SH-OPENED = '19920318'. APPEND SH. 
 
PROVIDE NAME AGE FROM SE 
		NAME	 FROM PR 
		*		FROM SH 
		BETWEEN BEGIN AND END. 
  ... 
ENDPROVIDE.

The three tables are processed according to the following schema:

ISE1 ISE2 ISE3
|-------| |-----------| |------------------------|
: : : : : :
: :IPR1 IPR2 : : IPR3 IPR4 :
: |----------|------| |--------------|------| :
: : : : : : : : : :
: : ISH1 : : : ISH2 : : :
|----------------------| |---------------------| :
: : : : : : : : : :
: : : : PROVIDE area : : :
|--------------------------------------------------|...
: : : : : : : : :
: : : : : : : : :
...|----|--|--|----|------| |--------------|------|...
result ranges

This PROVIDE loop is executed 7 times and produces the following sub-ranges:

In most of the loop passes, the fields SE_VALID, PR_VALID and SH_VALID contain 'X' . The exceptions to this are the 1st loop pass, where PR_VALID contains ' ', and the 3rd loop pass, where SE_VALID contains ' '.

Field contents (header lines) during the third loop pass:

SE-FROM = '01101991'
SE-TO = '04101991'
SE-NAME = ' '
SE-AGE = 0
PR-START = '01101991'
PR-END = '04101991'
PR-PRICE = 0
PR-NAME = 'Car'
SH-CLOSED = '04101991'
SH-STR = 'Gold Avenue'
SH-OPENED = '01101991'

Strictly speaking, if you imagine each range as a short way of writing a set of single values, this is an "outer join" of the tables.

Related

LOOP AT itab