Read Dataset CSV with Line Feeds in Cells - csv

Question

We are using the following code to read a CSV file from the Application Server:
OPEN DATASET file_str FOR INPUT IN TEXT MODE ENCODING DEFAULT.
*--------------------------------------------------*
* process and display output
*--------------------------------------------------*
DO.
CLEAR: lv_record,idat.
READ DATASET file_str INTO lv_record.
IF sy-subrc NE 0.
EXIT.
ELSE.
The problem we encounter now is that the CSV file holds Line Feeds in the cells:
If we read it with the above code the read dataset splits it in the middle of the cell instead of in the end.
What is the best way of handling this? We tried to read the file with the line feeds and do a replace all but we can't seem to visualize the line feeds in read dataset.
Thanks for your help!

Answer 1

This is a standard string handling issue - nothing specific to ABAP, you would encounter the same issue with BufferedReader.readLine(). Just check whether the line is complete (either contains the correct number of fields, or contains an even number of (un-quoted) cell-delimiters, i. e. "), and if it doesn't, read the next line and append it with CL_ABAP_CHAR_UTILITES=>CR_LF, then repeat.

Answer 2

This is the solution:
OPEN DATASET file_str FOR INPUT IN TEXT MODE ENCODING DEFAULT.
*--------------------------------------------------*
* process and display output
*--------------------------------------------------*
DATA: len TYPE i.
DATA: test TYPE string.
DATA: lv_new TYPE i,
lv_last_char TYPE c.
DATA: lv_concat TYPE string.
DO.
CLEAR: lv_record,idat, lv_concat.
READ DATASET file_str INTO lv_record.
IF sy-subrc NE 0.
EXIT.
ELSE.
"-- Get the string length
CALL FUNCTION 'STRING_LENGTH'
EXPORTING
string = lv_record
IMPORTING
length = lv_new.
"-- Check if the string is ended correctly
lv_new = lv_new - 1.
lv_last_char = lv_record+lv_new(1).
IF lv_last_char EQ '"'.
CONTINUE.
ELSE.
"-- Read next line
CONCATENATE lv_concat lv_record INTO lv_concat.
CLEAR lv_record.
WHILE lv_last_char NE '"'.
READ DATASET file_str INTO lv_record.
CALL FUNCTION 'STRING_LENGTH'
EXPORTING
string = lv_record
IMPORTING
length = lv_new.
lv_new = lv_new - 1.
lv_last_char = lv_record+lv_new(1).
CONCATENATE lv_concat lv_record INTO lv_concat.
ENDWHILE.
ENDIF.
IF lv_concat IS NOT INITIAL.
CLEAR lv_record.
MOVE lv_concat TO lv_record.
ENDIF.