Field lengths in unicode system.

Question

Hi gurus,
I am creating a dynamic internal table. To get the structure of DDIC table at run time i use the following code to create field catalog:
DATA : idetails TYPE abap_compdescr_tab,
         xdetails TYPE abap_compdescr.
DATA : xfc TYPE lvc_s_fcat,
         ifc TYPE lvc_t_fcat.
ref_table_des ?=
        cl_abap_typedescr=>describe_by_name( p_ztab_name ).
    idetails[] = ref_table_des->components[].
    LOOP AT idetails INTO xdetails.
      CLEAR xfc.
      xfc-fieldname = xdetails-name .
      xfc-datatype = xdetails-type_kind.
      xfc-inttype = xdetails-type_kind.
      xfc-intlen = xdetails-length.
      xfc-decimals = xdetails-decimals.
      APPEND xfc TO ifc.
    ENDLOOP.
The xdetails-length is doubled in a unicode system.
However it is correct as field length of table in non-unicode system.
I am then writing the data from internall table to file using
OPEN DATASET ld_fullpath FOR OUTPUT IN TEXT MODE ENCODING UTF-8
For ex. a numeric field in the file are written as 00 when its actual value in table is 0.
Please suggest how to solve this problem?
Should i build field catalog in some other way.
Assured points for helpful answers.
Regards,
Abhishek

Answer 1

Hi,
use this function module to get the filed info.
DATA:lta_dfies type dfies occurs 0 with header line.
CALL FUNCTION 'DDIF_FIELDINFO_GET'
    EXPORTING
      TABNAME              = TABLE
    TABLES
      DFIES_TAB            = lta_dfies.
loop at lta_dfies.
if (condition to get required).
perform build_fcat.
endif.
endloop.
form build_fcat.
CLEAR xfc.
xfc-fieldname = lta_dfies-FIELDNAME .
xfc-datatype = lta_dfies-DATATYPE.
xfc-inttype = lta_dfies-inttype.
xfc-intlen = lta_dfies-INTLEN.
xfc-decimals = lta_dfies-decimals.
APPEND xfc TO ifc.
endform.
rgds,
bharat.
Message was edited by:
        Bharat Kalagara