Issue with XML encoding

[asatish_kumar]

Dear SDN members,

I am facing somem issues with xml encoding for diacritical characters (special characters (non-english)).

For example below org unit name AfricaéçaçÃtest is converted to Africa a  test.
special chars are replaced with spaces , i used below code
IF lv_xml_data IS NOT INITIAL.
* ====================================================================================
*       Correct the Encoding of output
* ====================================================================================

lv_xml_data     TYPE string,
lv_buffer       TYPE xstring.

    EXPORT lv_xml_data TO DATA BUFFER lv_buffer.

    lo_conv = cl_abap_conv_in_ce=>create( input = lv_buffer
                                           encoding = '1105'
                                           ignore_cerr = 'X'
                                          replacement = '' ).

    lo_conv->convert( EXPORTING input = lv_buffer
                      IMPORTING data = lv_xml_data  ).

    SHIFT lv_xml_data UP TO '<root'.
  ENDIF.

  ev_xml_data = lv_xml_data.

Tried removing export which is causing the issue (AfricaéçaçÃtest  converts to Africaéçaçà test

i tried changing encoding types 4110 (utf-8) , 1160,1101,1100 etc. but it didn't work .

i tried this   CALL FUNCTION 'SCMS_STRING_TO_XSTRING'   EXPORTING  text   = lv_xml_data    IMPORTING  buffer = lv_buffer.

and also string to binary and then binary to xstring but none has slved my problem .

Appretiate your inputs to resolve ths issue .

Thanks & Regards
Satish

[former_member182354]

Hi Satish,

               CALL FUNCTION 'SCMS_STRING_TO_XSTRING this FM has this issue.

               Use BCS class to avoid this where the special characters are sent without truncation or dump.

      try.
call method cl_bcs_convert=>string_to_solix
EXPORTING
iv_string   = lv_string
iv_codepage = lc_codepage
iv_add_bom  = gc_x
IMPORTING
et_solix    = lt_solix.
catch cx_bcs .
endtry.

*-- Create persistent send request
l_send_request = cl_bcs=>create_persistent( ).
wt_contents[] = t_contents[].

*-- Get the length of the Document
describe table wt_contents lines l_cnt.
read table wt_contents into ws_contents index l_cnt.
l_doc_len = ( l_cnt - 1 ) * 255 + strlen( ws_contents ).
*-- Subject of the mail
l_sub = w_mail_subj.
*-- Create Document
try.
l_document = cl_document_bcs=>create_document(
i_type       = lc_htm
i_text       = wt_contents
i_length     = l_doc_len
i_subject    = l_sub
  i_language   = sy-langu
i_importance = '1' ).
catch CX_DOCUMENT_BCS.
endtry.
*-- Subject of the mail
move w_mail_subj to l_subj.
w_document = l_document.

Raghav

[RaymondGiuseppi]

Look in Code Gallery for Escape HTML.

Regards,

Raymond

[Former Member]

You can use class cl_abap_conv_x2x_ce to change encoding of hex representation of text.

Looking at your example, you are trying to change encoding of a text directly.

Consider é, Hex representation for UTF-8 encoding would be C3A9. When this hex data is interpreted as latin encoded text, text would be é.

See below snippet that is converting AfricaéçaçÃtest to Africaéçaç#test.

1. TRY .
2.     DATA: text    TYPE string VALUE 'AfricaéçaçÃtest',
3.           buffer TYPE xstring,
4.           conv    TYPE REF TO cl_abap_conv_out_ce.
6.     conv = cl_abap_conv_out_ce=>create(
7.              encoding = '1164'
8.              ignore_cerr = 'X'  ).
10.     conv->convert( EXPORTING data = text
11.                    IMPORTING buffer = buffer ).
13.     WRITE / buffer.
14.     CLEAR text.
15.     DATA conv2 TYPE REF TO cl_abap_conv_in_ce.
17.     conv2 = cl_abap_conv_in_ce=>create(
18.               encoding = 'UTF-8'
19.               ignore_cerr = 'X'   ).
21.     conv2->convert(
22.       EXPORTING input = buffer
23.       IMPORTING data = text ).
25.     WRITE / text.
26.   CATCH cx_root.
27. ENDTRY.

/.

[asatish_kumar]

Hi Manish ,

The above code is replacing accent chars with # , i want without # .

please suggest.

Thanks & Regards

Satish

[Former Member]

It is supposed to show #, as à is not followed by something that can directly be interpreted correctly in UTF-8. You can dive deeper into hex level and do some substitutions that correct the output. Hex equivalent of Ãtest is C374657374. C3 could not be converted, so you get a #. In order to get à in UTF-8, C3 should be replaced by C383.

So, do not ignore the conversion error, and apply suitable substitution at hex level.