REDUCE not concatenating strings

[MarcC]

Why does the following code with REDUCE not concatenate the table into the string 'ABC'?

DATA t TYPE TABLE OF kna1.
t = VALUE #(
  ( kunnr = 'A' )
  ( kunnr = 'B' )
  ( kunnr = 'C' )
).

DATA(log1) = REDUCE string( INIT x1 = '' FOR wa IN t NEXT x1 = x1 && wa-kunnr ).

WRITE: /, log1. " A

DATA log2 TYPE string.
LOOP AT t INTO DATA(wa2).
  log2 = log2 && wa2-kunnr.
ENDLOOP.

WRITE: /, log2. " ABC

The LOOP approach works just fine.

[MarcC]

Why does the following code with REDUCE not concatenate the table into the string 'ABC'?

DATA t TYPE TABLE OF kna1.
t = VALUE #(
  ( kunnr = 'A' )
  ( kunnr = 'B' )
  ( kunnr = 'C' )
).

DATA(log1) = REDUCE string( INIT x1 = '' FOR wa IN t NEXT x1 = x1 && wa-kunnr ).

WRITE: /, log1. " A

DATA log2 TYPE string.
LOOP AT t INTO DATA(wa2).
  log2 = log2 && wa2-kunnr.
ENDLOOP.

WRITE: /, log2. " ABC

The LOOP approach works just fine.

[MarcC]

OK, I discovered it is because x1 was implicitly char(1) and could contain only one character. Here is the corrected code:

DATA t TYPE TABLE OF kna1.
t = VALUE #(
  ( kunnr = 'A' )
  ( kunnr = 'B' )
  ( kunnr = 'C' )
).

DATA(log1) = REDUCE string( INIT x1 = CONV STRING( '' ) FOR wa IN t NEXT x1 = x1 && wa-kunnr ).

WRITE: /, log1. " ABC

[MBartsch71]

You can achieve the same by sing backticks ` instead of single quotes '.

[Raifloresca]

Hi,

You can declare the type of the X1 variable inside the REDUCE statement. That's what I would do. More details can be found in the REDUCE documentation

Regards.