SAPFree generatorPDF, Excel, Word exportSAP ECC / S/4HANA

ABAP Internal Table Declaration Generator

Choosing between standard, sorted and hashed internal tables decides the speed of every READ TABLE in your program. This generator writes the declaration for the kind and key you pick, a matching local type, and the READ, LOOP and DELETE statements that go with it.

Fill in

Table variable name

Line type (DDIC structure)

Table kind

Key kind

HASHED tables must be UNIQUE. STANDARD tables must be NON-UNIQUE.

Key component(s)

Space separated field names.

Lookup value variable

Generated code

Download:

1. Declaration with a DDIC line type

DATA lt_items TYPE SORTED TABLE OF mara
  WITH UNIQUE KEY matnr.

2. Declaration with a local line type

TYPES: BEGIN OF ty_line,
         matnr TYPE matnr,
         maktx TYPE maktx,
         menge TYPE menge_d,
       END OF ty_line.

DATA lt_items TYPE SORTED TABLE OF ty_line
  WITH UNIQUE KEY matnr.

3. Common operations

* Read one row by key
READ TABLE lt_items ASSIGNING FIELD-SYMBOL(<ls_row>)
  WITH KEY matnr = lv_matnr.
IF sy-subrc = 0.
* <ls_row> points at the row, changes are written to the table
ENDIF.

* Loop without copying rows
LOOP AT lt_items ASSIGNING FIELD-SYMBOL(<ls_item>).
* use <ls_item>-field here
ENDLOOP.

* Delete by key
DELETE lt_items WHERE matnr = lv_matnr.

* Standard tables only: sort and remove duplicates
* SORT lt_items BY matnr.
* DELETE ADJACENT DUPLICATES FROM lt_items COMPARING matnr.

Rules the compiler enforces: HASHED tables need a UNIQUE key, STANDARD tables only allow NON-UNIQUE, and SORT and DELETE ADJACENT DUPLICATES are meant for standard tables. Replace the sample local type with your own components. Inline FIELD-SYMBOL(...) needs ABAP 7.40 SP02 or later.

Review a performance-sensitive program

Send us a slow custom program and we will describe where the table access pattern could improve.

Optional. The tool and every export work without this. No spam.

How to use it

Pick the variable name, the DDIC structure or your own line type, the table kind and the key fields. The first block gives the one-line DATA declaration. The second block shows the same table with a local TYPES structure, which is what you use when only some fields are needed. The third block gives the statements you will write next. Copy what you need, rename the lookup variable and replace the example components with your own.

Standard, sorted or hashed: which to choose

A standard table is read by linear search unless you SORT it and use BINARY SEARCH, so it suits small tables and sequential processing. A sorted table keeps rows in key order and uses binary search on the key, which is the best default when you need ordered access and partial key reads. A hashed table gives constant time access on the full unique key, which is best for large lookup tables read by exact key and never processed in order. Choose by the access pattern, not by habit.

Standard: append and loop, small lookups
Sorted: ordered reads and range access by key prefix
Hashed: large tables read only by full unique key

Field symbols, references and copies

LOOP AT with ASSIGNING gives a pointer to the table row, so there is no copy and changes write straight into the table. Use INTO when you want a work area copy. For wide rows, such as full MARA lines, ASSIGNING is noticeably cheaper. READ TABLE with WITH KEY on a sorted or hashed table is optimised when the key fields match the table key, so use WITH TABLE KEY if you want the compiler to insist on that.

Common mistakes

Frequent errors are reading a big standard table inside a loop without sorting and binary search, declaring a hashed table with a non-unique key, expecting insertion order from a sorted table, and deleting rows while looping with an index. Also check sy-subrc after every READ TABLE, because an unchecked failed read leaves the previous field symbol assignment untouched in older code.

Frequently Asked Questions

Which internal table type is fastest?

It depends on how you read it. A hashed table is fastest for exact lookups on the full unique key. A sorted table is fast for key or partial-key reads and ordered loops. A standard table is fine for small data or when you read it once from top to bottom.

Why do I get a syntax error for UNIQUE KEY on a standard table?

Standard tables have no uniqueness check because rows are appended without lookup, so only a NON-UNIQUE key is allowed. If you need uniqueness, use a sorted or hashed table, or remove duplicates yourself with DELETE ADJACENT DUPLICATES after sorting.

What does WITH DEFAULT KEY do?

It uses all character-like and byte-like components as the key. It is common in older code, but it is not recommended for new code, and the SAP code inspector flags it. Name the key fields you actually need instead.

Should I use ASSIGNING or INTO in a LOOP?

ASSIGNING avoids copying the row and allows direct changes, so it is usually better for wide rows. INTO copies each row into a work area, which is safer if you do not want to modify the table. Both are valid, choose on intent and row width.

If your ABAP is mostly moving ERP data into tables for reports, we can show how a governed read-only layer answers the same questions without new development.

Talk to an Expert Explore More Free Tools

Related Tools

SAP

More SAP tools

See all

Back to All Tools