SAP ABAP Internal Table: Create, Read, Populate, Copy & Delete

By : Sophia Mitchell Sophia Mitchell
Hours Updated

⚡ Smart Summary

SAP ABAP Internal Table objects hold data from a fixed structure for dynamic use inside a program. This page explains work areas, header lines, the four creation methods, and the statements that populate, copy, read, and delete table rows.

  • 📋 Core Definition: An internal table stores rows of identical structure in main memory, most often to format data read from a database table.
  • 🧾 Work Area Role: A work area is a single row with the same format as the table, and it processes the table one line at a time.
  • 🏷️ Header Line: A table with a header line receives an implicit work area, so rows can be inserted or read directly without naming one.
  • 🛠️ Four Creation Methods: TYPES statement, reference to an existing table, reference to an existing structure, or a new structure declared with OCCURS.
  • ➕ Populate: APPEND adds one row, COLLECT aggregates numeric fields for a duplicate key, and INSERT places a row at a chosen index.
  • 🔍 Read and Copy: LOOP AT reads every row, READ TABLE reads one row by index, and MOVE or the equals sign copies a whole table.
  • 🗑️ Delete: DELETE inside a LOOP removes the current row, and DELETE … INDEX removes the row at a known position.
Read More

SAP ABAP Internal Tables

What is an Internal Table?

INTERNAL TABLE are used to obtain data from a fixed structure for dynamic use in ABAP. Each line in the internal table has the same field structure. The main use for internal tables is for storing and formatting data from a database table within a program.

An internal table always works together with a work area, so that concept comes next.

What is a Work Area?

Work areas are single rows of data. They should have the same format as any of the internal tables. It is used to process the data in an internal table one line at a time.

Difference Between Internal Table and a Work Area?

A picture says a thousand words

Difference Between Internal Table and a Work Area

The diagram shows the table holding many rows while the work area holds only the row currently being processed. Whether that work area is created by the system or by the developer depends on the table type described below.

Types of Internal Tables

There are two types of internal tables.

  1. Internal tables with HEADER line
  2. Internal tables without HEADER line.

Internal Tables with Header Line

  • Here the system automatically creates the work area.
  • The work area has the same data type as internal table.
  • This work area is called the HEADER line.
  • It is here that all the changes or any of the action on the contents of the table are done. As a result of this, records can be directly inserted into the table or accessed from the internal table directly.

Internal Tables without Header Line :

  • Here there is no work area associated with the table.
  • Work area is to be explicitly specified when we need to access such tables.
  • Hence these tables cannot be accessed directly.

💡 Tip: A table with a header line carries the same name for the table and its work area, which makes the code ambiguous. New programs should declare a table without a header line and pass an explicit work area.

Standard, Sorted, and Hashed Internal Tables

Besides the header line distinction, every internal table also has a table kind that decides how a row is found. The kind is chosen with the TYPE STANDARD TABLE, TYPE SORTED TABLE, or TYPE HASHED TABLE addition, and it has a direct effect on runtime.

Table kind Access method Key Best used for
Standard Index and key, with a linear search on the key Non unique Sequential processing, and small tables read with an index
Sorted Index and key, with a binary search on the key Unique or non unique, kept sorted automatically Larger tables read repeatedly through the leading key fields
Hashed Key only, with a hash algorithm Unique Very large tables that are always read by the full key, such as a buffer of database records

A standard table is the default and suits sequential processing. A hashed table keeps the access time constant no matter how many rows are stored, which is why the performance rules of Open SQL recommend it for buffering records.

Creating Internal Tables

There are many ways to create an Internal Table. Lets look at them one by one-

1.By Using the Type Statement

Let us now create a Internal table itab using the TYPE statement.

The syntax is –

Types : begin of line,

column1 type I,

column2 type I,

end of line.

Example:

TYPES : begin of line,

empno		type I,

empname(20)   	type c	,

end of line.

The TYPES statement creates a structure line as defined.

To actually create an Internal Table itab use the following command-

Data itab type line occurs 10.

An internal table itab is created with the structure of line.Besides declaring the structure of an internal table, the OCCURS clause also defines how many table entries are maintained in main storage(in this case 10). Extra records are written out to paging area and can effect performance

2.By referring to another Table

You can create an internal table by referring to an existing table. The existing table could be a standard SAP table, a Z table or another internal table.

Syntax-

Data <f> <type> [with header line].

Example-

DATA itab TYPE line OCCURS 10 with header line.

Here an internal table itab is created of the type line with a header line. Please note “with header line” is optional.

3.By referring to existing Structure

Syntax-

Data	<f> LIKE <struct> occurs n [with header line].

Example-

DATA itab LIKE sline OCCURS 10.

Here a table itab is created having a structure same as that of sline

4.By creating a new Structure

Let us now create an internal table with a structure of our own. Here the table is created with an Header line, by default.

Syntax –

Data : Begin of <f> occurs <n>,

<component declaration>,

................................,

End of <f>.

Example –

Data : Begin of itab occurs 10,

column1       type I,

column2(4)  type C,

column3      like  mara-ernam,

End of itab.

Internal table itab is created

An empty table is of little use, so the next section fills it with rows.

RELATED ARTICLES

Populating Internal Tables

Now that we have successfully created some internal tables, let us see how do we populate them with some records. There are various methods available to populate tables

1.Append Data line by line

The first method available is the use of the APPEND statement.

Using the APPEND statement we can either add one line from another work area to the internal table or we can add one initial line to the internal table..

Syntax –

APPEND [<wa> TO / INITIAL LINE TO] <itable>.

Here work area <wa> or the Initial Line is appended to the internal table <itable>.

The system variable SY-TABIX contains the index of the appended line.

Example:

Data: Begin of itab occurs 10,

col1 type C,

col2 type I,

end of itab.

Append initial line to itab.

Results : ‘ ‘ ‘0’

Initial lines adds a line initialized with the correct value for its type to the table. Here , col1 is an character and col2 is a integer. Then APPEND initial line, adds a line initialized with respect to the data type of the columns, i.e. space for col1 and 0 for col2.

2.Using COLLECT statement

COLLECT is another form of statement used for populating the internal tables. Generally COLLECT is used while inserting lines into an internal table with unique standard key.

Syntax-

COLLECT [<wa> INTO] <itable>.

Incase of tables with Header line, INTO option is omitted. Suppose there is already an entry having a key same as the one you are trying to append, then a new line is not added to the table, but the numeric fields of both the entries are added and only one entry corresponding to the key is present. Value of SY-TABIX is changed to the row of the original entry. Else COLLECT acts similar to APPEND and SY-TABIX contains the index of the processed line.

3.Using INSERT statement

INSERT statement adds a line/work area to the internal table. You can specify the position at which the new line is to be added by using the INDEX clause with the INSERT statement.

Syntax

INSERT [<wa> INTO / INITIAL LINE INTO] <itable> [index <idx>].

Here, the work area <wa> or INITIAL LINE is inserted into internal table <itable> at index <idx>.

Copying Internal Tables

The contents of one internal table can be copied to another by using the APPEND LINES or INSERT LINES statement. A more simpler way is to use any of the following syntax’s.

MOVE  <itab1> To <itab2>.

OR

<itab1> = <itab2>.

These copy the contents of ITAB1 to ITAB2. Incase of internal tables with header line we have to use [] inorder to distinguish from work area. So, to copy contents of internal tables with header line the syntax becomes,

itab1[] = itab2[].

Read Internal Tables

We are now familiar with the creation of internal tables and populating them with data. We will now see how do we actually use the data or retrieve the data from the internal tables.

1. Using Loop -Endloop

One of the ways of accessing or reading the internal table is by using LOOP-ENDLOOP.

Syntax

LOOP AT <itable> [INTO <wa>]

...................................

ENDLOOP.

Here when you say LOOP AT ITABLE, then the internal table ITABLE is read line by line. You can access the values of the columns for that line during any part of the LOOP-ENDLOOP structure. The value of the SY-SUBRC is set to 0, even if only one record is read.

2. Using READ

The other method of reading the internal table is by using the READ statement.

Syntax-

READ TABLE <itable> [INTO <wa>] INDEX <idx>.

This statement reads the current line or line as specified by index <idx>. The value of SY-TABIX is the index of the line read. If an entry with the specified index is found, then SY-SUBRC is set to 0. If the specified index is less than 0, then run-time error occurs. If the specified index exceeds table size then SY-SUBRC is set to 4.

⚠️ Warning: READ TABLE … INDEX 0 raises a runtime error, so an index calculated at runtime should be checked before the statement runs.

Deleting Internal Tables

There are many ways for deleting lines from an internal table.

1.Deleting lines in a loop.

This is the simplest way for deleting lines.

Syntax

DELETE <ITABLE>.

This statement works only within a loop. It deletes the current line. You can delete the lines in a loop conditionally by adding the WHERE clause.

2.Deleting lines using the index.

This is used to delete a line from internal table at any know index.

Syntax

DELETE <ITABLE> INDEX <IDX>.

The line with the index <IDX> is deleted. The index of the following line is decremented by 1.

3.Deleting duplicate lines.

Duplicate rows are removed after the table has been sorted by the key fields.

SORT itab BY col1.

DELETE ADJACENT DUPLICATES FROM itab COMPARING col1.

Only adjacent rows are compared, which is why the SORT statement has to run first. To empty the table completely, use CLEAR itab[] or REFRESH itab.

FAQs

CLEAR itab empties the header line only, and CLEAR itab[] empties the body. REFRESH always empties the body but keeps the memory reserved. FREE empties the body and releases the memory as well.

SY-TABIX holds the row number of the internal table currently being processed. SY-INDEX holds the pass number of a DO or WHILE loop. A hashed table leaves SY-TABIX at zero, because it has no index.

LOOP AT itab ASSIGNING a field symbol avoids copying every row into a work area, so it runs faster on large tables and allows the row to be changed in place without a MODIFY statement.

Yes. AI assistants in ABAP development tools rewrite OCCURS and header line declarations into TYPE STANDARD TABLE OF with an explicit work area, and they flag the statements that break once the header line disappears.

AI performance advisors inspect how a table is filled and read, then recommend a sorted or hashed table when a linear search dominates the runtime. The recommendation is confirmed with a trace in transaction SAT.

Summarize this post with: