ABAP Keyword Documentation → ABAP - Reference → Processing Internal Data → Internal Tables → Processing Statements for Internal Tables → LOOP AT itab

LOOP AT itab - cond

Short Reference

Other versions: 7.31 | 7.40 | 7.54

Syntax


 ... [USING KEY keyname] [FROM idx1] [TO idx2] [WHERE log_exp|(cond_syntax)] ... .

Extras

1. ... USING KEY keyname

2. ... [FROM idx1] [TO idx2]

3. ... WHERE log_exp

4. ... WHERE (cond_syntax)

Effect

USING KEY keyname is used to determine the table key with which the loop is executed. The table rows to be read in a LOOP loop can also be limited using optional conditions. If no conditions are declared, all table rows are read.

Within the loop, the key being used can be addressed using the predefined loop_key. This is possible in all statements where the table key keyname is used and where it can be declared explicitly. This type of statement must then be executed in the loop itself. Including the statement in a procedure that is called in the loop is not sufficient.

Addition 1

... USING KEY keyname

Effect

The USING KEY addition can be used to specify a table key keyname with which the processing is carried out. The specified table key influences the sequence in which the table rows are accessed, and the evaluation of the remaining conditions.

If the primary table key is specified using the name primary_key, the processing behaves in the same way as when no key is explicitly specified. If a secondary table key is specified, the sequence in which the rows are accessed is as follows:

Specification of a sorted key
The rows are processed by ascending row number in the secondary table index In each loop pass, the system field sy-tabix contains the row number of the current row in the associated secondary table index.
Specification of a hash key
The rows are processed in the sequence in which they were inserted into the table. In each loop pass, the system field sy-tabix contains the value 0.

Notes

Unlike the processing of a hash table when a primary key is used, a preceding sort using the SORT statement has no influence on the processing sequence when a secondary hash key is specified.
If a secondary table key is specified, any WHERE condition also specified must be optimizable. Otherwise a syntax error occurs or an exception is raised.

Example

See Loop though internal table with key specification.

Addition 2

... [FROM idx1] [TO idx2]

Effect

These additions have the effect that only table rows from row number idx1, or up to row number idx2, are taken into account in the table index used. If only FROM is specified, all rows of the table from row number idx1 up to and including the last row are taken into account. If only TO is specified, all rows in the table from the first row up to row number idx2 are taken into account.

If the addition USING KEY is not used, or the primary table key is specified in keyname, the additions FROM and TO can only be used for index tables. In this case, they refer to the row numbers of the primary table index.

If a sorted secondary key is specified in keyname after USING KEY, the additions FROM and TO can be used for all table types and refer to the row numbers of the secondary table index.

idx1 and idx2 are numerical expression positions of operand type i. The following limitations apply:

If the value of idx1 is less than 0, it is set to 1 in the LOOP statement and causes a runtime error in every other statement. If the value of idx1 is greater than the total number of table rows, no processing takes place.
If the value of idx2 is less than 0, the LOOP statement is not carried out and in every other statement it leads to a runtime error. If the value of idx2 is greater than the number of table rows, it is set to the number of table rows.
If the value of idx2 is less than the value of idx1, no processing takes place.

The value of idx1 is evaluated once when the loop is entered. Any changes to idx1 during loop processing are not taken into account. In contrast, the value of idx2 is evaluated in each loop pass and any changes made to idx2 during loop processing are taken into account.

Note

To determine when loop processing is exited and whether the value specified in idx2 has been reached, the current line number is evaluated. Note that this number can be changed if lines are inserted or deleted during a loop pass as described in LOOP. As a result, there may be more loop passes (if lines are inserted) or fewer loop passes (if lines are deleted) than is specified by the difference between idx2 and idx1.

Addition 3

... WHERE log_exp

Effect

Static WHERE condition. All rows are processed for which the condition after WHERE is met. You can specify WHERE for all table categories.

You can specify a logical expression log_exp after WHERE in which the first operand of each operation is a component of the internal table. You cannot specify a predicate function. The components of the internal table must be specified as individual operands and not as part of an expression. You cannot use parenthesized character-like data objects to specify a component dynamically here. The remaining operands of a comparison are general expression positions at which any suitable inidividual operands or expressions can be specified, but no components of the internal table. All logical expressions are possible except IS ASSIGNED, IS SUPPLIED, and the obsolete IS REQUESTED. The specified components can have any data type. The relevant comparison rules apply to the evaluation.

When standard tables are accessed without a secondary key being specified, the access is not optimized. This means that all rows of the internal table are tested for the logical expression of the WHERE addition.

When using a sorted key or a hash key (that is, when accessing a sorted table, a hashed table, or a secondary table key), an attempt is made to optimize the access as described under Optimization of the WHERE Condition. If the corresponding prerequisites are not met:

the logical expression cannot be transformed to a key access

the logical expression has the same result as the resulting key access

then no optimization takes place when a sorted table or a hashed table is accessed using the primary table key. Any access using a secondary table key produces a syntax error or exception. If the remaining prerequisites for the optimization are met, then the static WHERE condition cannot contain any duplicate or overlapping key specifications.

Notes

When using a WHERE condition, note that the comparison rules for incompatible data types apply when comparing incompatible data objects. Here, the data types involved determine which operand is converted. If the additions WITH TABLE KEY and WITH KEY of the statement READ are used, however, the content of the specified data objects is always converted to the data type of the columns before the comparison. This can produce varying results.
If possible, all operands of the logical expression should be in compatible pairs, so enabling the WHERE condition to be optimized.
If a selection table is specified after IN as a logical expression, note that the expression at the initial table is always true and then all rows are edited.
The logical expression declared after WHERE is evaluated once when the loop is entered. Any changes to the second operand during loop processing are ignored.

Example

The following example demonstrates the differences in behavior of a WHERE condition and a key access with WITH TABLE KEY. With LOOP AT itab WHERE, the rule for the comparison of character-like data types applies. The short column content "AA" is first filled with spaces to change the length to 4. It is then compared to "AAXX". No matching row is found. With READ TABLE itab WITH TABLE KEY, the content of text_long is converted to the value "AA" before the comparison, by truncating two characters, and then compared to the column content. The result is output without errors.

DATA text_short TYPE c LENGTH 2. 
DATA text_long  TYPE c LENGTH 4. 

DATA itab LIKE TABLE OF text_short WITH NON-UNIQUE KEY table_line. 

text_short = 'AA'. 
text_long  = 'AAXX'. 

APPEND text_short TO itab. 

LOOP AT itab INTO text_short WHERE table_line = text_long. 
ENDLOOP. 
WRITE: / 'LOOP:', sy-subrc. 

READ TABLE itab INTO text_short WITH TABLE KEY table_line = text_long. 
WRITE: / 'READ:', sy-subrc.

Addition 4

... WHERE (cond_syntax)

Effect

Dynamic WHERE condition. You can declare a character-type data object or standard table with character-type data type for cond_syntax. This data object or standard table is initial or contains the syntax of a logical expression (in accordance with the rules of the static WHERE condition) when the statement is executed.

As in ABAP Editor, the syntax in cond_syntax is not case-sensitive. When you declare an internal table, you can distribute the syntax across multiple rows. If cond_syntax is initial when the statement is executed, the logical expression is true. An invalid logical expression raises an exception of the class CX_SY_ITAB_DYN_LOOP.

The obsolete logical operators (><, =>, and =<) are not supported by cond_syntax.

Notes

Of the calculation expressions, only arithmetic expressions are currently supported. The string expressions and bit expressions are not supported. You cannot specify string functions or bit functions currently either.
For optimization reasons, the dynamic WHERE condition is not evaluated for empty tables. This means that an exception is not raised for empty internal tables if the logical expression contains an error.

Example

Reading lines with particular line numbers in the primary table index that meet a condition. The static and dynamic declaration of a WHERE condition are shown.

DATA: BEGIN OF line, 
         col1 TYPE i, 
         col2 TYPE i, 
      END OF line. 

DATA itab LIKE SORTED TABLE OF line WITH UNIQUE KEY table_line. 

DATA num TYPE i VALUE 400. 
DATA dref TYPE REF TO i. 
DATA cond TYPE string. 

DO 30 TIMES. 
  line-col1 = sy-index. 
  line-col2 = sy-index ** 2. 
  APPEND line TO itab. 
ENDDO. 

GET REFERENCE OF num INTO dref. 

LOOP AT itab INTO line FROM 10 TO 25 WHERE col2 > dref->*. 
  WRITE / line-col2. 
ENDLOOP. 

SKIP. 

cond = 'col2 > dref->*'. 

LOOP AT itab INTO line FROM 10 TO 25 WHERE (cond). 
  WRITE / line-col2. 
ENDLOOP.