Markup not displaying Doc/docx/PDF properly

Question

Hey everyone,
I'm using Oracle text in conjunction with Apex, all on top of the 11G database. My issue is that using the Oracle text Markup, many files types do not display properly. The level of wrongness varies, from some readable text to complete gibberish. This happens to all version of PDFs, Docs, and Docx files. A sample of what appears is below:
Æ
@à°€P ðÀ!dðgd£n=gd8h²„h^„hgd8h²^
&F
Æ
p@à°€P ðÀ!dðEÆ€^ZfFG·ðgd÷i+
&F„˜þ`„˜þgd÷i+
&F
Æ
pà°€P ðÀ!dðgdýOôHN”³´ÔÕûý     "#f‘ðáÕáÁ²Õ² Ž|n]L²8-h£n=h£n=CJaJ&h£n=h÷i+6B*CJ\]aJph h8h²h÷i+5B*CJaJph h£n=h÷i+5B*CJaJphh£n=5B*CJaJph#h£n=h£n=5>*B*CJaJph#h£n=h÷i+5>*B*CJaJph#h8h²h÷i+5>*B*CJaJphh8h²h÷i+B*CJaJph&h8h²h÷i+6B*CJ\]aJphh£n=B*CJaJphh£n=h÷i+B*CJaJphh£n=h£n=B*CJaJph´üý#â„_B
&F
However, using a simple .txt file, the result is nearly perfect and everything is complete readable.
create or replace PROCEDURE test
     (p_id      IN VARCHAR2,
       p_query   IN VARCHAR2)
    AS
      v_clob         CLOB;
      v_read_amount INTEGER := 32767;
      v_read_offset INTEGER := 1;
      v_buffer      VARCHAR2 (32767);
    BEGIN
        htp.p('<b>HTML version with highlighted terms</b>');
     CTX_DOC.MARKUP
       (index_name => 'DOCUMENTS_INDEX',
         textkey    => p_id,
         text_query => p_query,
         restab     => v_clob,
         starttag   => '<i><font color=red>',
         endtag     => '</font></i>');
     LOOP
        BEGIN
            dbms_lob.read(v_clob,v_read_amount,v_read_offset,v_buffer);
            htp.p(v_buffer);
            v_read_offset := v_read_offset + v_read_amount;
            v_read_amount := 32767;
        EXCEPTION
          WHEN NO_DATA_FOUND THEN EXIT;
        END;
     END LOOP;
   END test;The above is the procedure I'm using if that helps. If anyone has ideas about what's going wrong, I'd greatly appreciate it.

Answer 1

Well, I figured this one out on my own. I had been using a multi-column datastore, and it turns out that you have to turn on auto-filtering for each column you add. Once I had done this and rebuilt my index, my markup was much more accurate and readable. Hopefully this can help out anyone else going through this issue.