CX_SY_CONVERSION_CODEPAGE - Error when reading from file in unix directory

Question

Hi
We have a custom report which reads data from a file put in the unix directory. This report has been running for good 3 years now without any code changes.
For a few days the report is dumping with the following error
CX_SY_CONVERSION_CODEPAGE
Short text
    Character set conversion is not possible.
What happened?
    When converting a text from code page '4110' to code page '4102',
    the system either
    - encountered a character which cannot be represented in one of
      the two code pages; or
    - found that this type of conversion is not supported.
The code where the dump occurs is:
load the file from the given path
  OPEN DATASET  p_p_lfdata FOR INPUT IN TEXT MODE ENCODING DEFAULT WITH SMART LINEFEED.
  IF sy-subrc EQ 0.
    DO.
>>>>>>>>      READ DATASET  p_p_lfdata INTO  l_rec. " Dump occurs here
      IF sy-subrc = 0.
I understand this error is thrown in the OPEN DATASET statements in case of non-unicode compliance, but this statement does not throw an error while the READ DATASET statement does..... Any particular reason why the READ statement is failing even though OPEN is successful???

Answer 1

Hi,
you open the according file in DEFAULT mode. I assume you do this on a Unicode system.
Then the program assumes that the file is encoded in UTF-8 (SAP number 4110).
Now it seems that the file contains some characters, which are not encoded in UTF-8.
I would propose to investigate the code page, on which the file is created.
Ideally this should be UTF-8. If this is not the case, try to change it.
If this is not possible, then adapt the ABAP program (e.g. use Non-Unicode mode).
Best regards,
Nils Buerckel