topic Truncate string based on byte in Application Development and Automation Discussions

Truncate string based on byte

former_member227603 — Sun, 28 Feb 2010 05:24:09 GMT

Hi Experts,

Looking for help.

I have a requirement where i need to truncate string based on byte count.

As some of chinese or other language char are multibyte and in string i need to truncate a part based on no of byte.

I know in normal scenario we do it as STRING +offset(length) but that gives in character mode count.

But based on no off byte i have to extract a substring.

Hope i am clear about the requirement.

Regards,

Digvijay Singh

Re: Truncate string based on byte

Former Member — Sun, 28 Feb 2010 08:54:46 GMT

Hi,

I think if you are working in ECC version, which is Unicode, it's the same for you to use the old way.

Cheers,

Edited by: NI SHILIANG on Feb 28, 2010 9:55 AM

Re: Truncate string based on byte

Former Member — Mon, 01 Mar 2010 08:09:02 GMT

Hi Digvijay,

Not sure if I'm missing the simple solution (and if I understand the problem, because usually you actually want to handle characters and <i>not bytes</i>), but here's a fairly convoluted one for truncating a string at a specific byte count:


  CONSTANTS:
    co_input  TYPE string VALUE 'u7E41u4F53u5B57'.
  DATA:
    l_codepage TYPE cpcodepage,
    l_encoding TYPE abap_encoding,
    l_conv_out TYPE REF TO cl_abap_conv_out_ce,
    l_conv_obj TYPE REF TO cl_abap_conv_obj,
    l_string  TYPE string,
    l_xstring TYPE xstring.

  CALL FUNCTION 'SCP_CODEPAGE_FOR_LANGUAGE'
    EXPORTING
      language = sy-langu
    IMPORTING
      codepage = l_codepage
    EXCEPTIONS
      OTHERS   = 0.

  l_encoding = l_codepage.
  l_conv_out = cl_abap_conv_out_ce=>create( encoding = l_encoding ).
  l_conv_out->convert( EXPORTING data   = co_input
                       IMPORTING buffer = l_xstring ).

  sy-fdpos = xstrlen( l_xstring ) - 3.
  l_xstring = l_xstring(sy-fdpos).

  CREATE OBJECT l_conv_obj.
  l_conv_obj->convert( EXPORTING inbuff    = l_xstring
                                 outbufflg = 0
                       IMPORTING outbuff   = l_string ).

  WRITE: / 'Before:', co_input, ';  After:', l_string.

When I run this example in our system (codepage 4103, which corresponds to utf-16 little endian) I get the following output:


Before: 繁体字 ;  After: 繁

The convoluted coding is in my opinion necessary to avoid splitting multi-byte characters in half. So in the example you can see that I take off three bytes, but it actually results in omitting the last two characters (as we basically got one and a half). I'd hope that somebody has a shorter and more elegant solution, let's see...

If you'd know that your strings never contain [surrogate pairs|http://unicode.org/faq/utf_bom.html#utf16-2] you'd have it much simpler though, because then you'd know that one character corresponds to exactly two bytes on the application server (since application server always uses UTF16).

Cheers, harald

Re: Truncate string based on byte

Sandra_Rossi — Mon, 01 Mar 2010 12:45:44 GMT

I guess you ask the question because you are in a NON-unicode system. As explained in [ABAP documentation - Conversion Table for Source Field Type c|http://help.sap.com/abapdocu_70/en/ABENCONVERSION_TYPE_C.htm], you may call method CL_SCP_LINEBREAK_UTIL=>STRING_SPLIT_AT_POSITION, to be used especially for non-unicode double-byte characters.