Is it possible to avoid "BY INDEX ROWID" access?

Question

Hi,
I'm running a query for the primary key of a table that has a text index on a text column, something like
SELECT pkey FROM table WHERE CONTAINS( text, 'foobar' ) > 0.
As expected, the index is scanned for the matching ROWIDs and then the rows are retrieved ( BY INDEX ROWID ).
I wonder if its possible to include the primary key column in the text index, so that the table doesn't have to be access at all when querying for the primary key ( this is possible with regular indexes ).
I tried
1. creating a Composite Domain Index ( i.e. including the primary key column in the FILTER BY clause of the index )
2. creating a SDATA section on pkey
3. creating a MDATA section on pkey
In all cases, the rows were still accessed by ROWID.
Any ideas?

Answer 1

Our index is called "docidx". We are looking for the term "oracle". We want a count of the total hits, and a hitlist of the first two hits. For each hit we want to fetch back "author" and "pubDate" sdata sections, as well as the rowid.
We also want some summary information: for pubDate and author sections we want a count for each unique value in each section.
ctx_query.result_set('docidx', 'oracle',
    '<ctx_result_set_descriptor>
       <count/>
         <hitlist start_hit_num="1" end_hit_num="2" order="pubDate desc, score desc">
           <score/> <rowid/>
           <sdata name="author"/>
           <sdata name="pubDate"/>
         </hitlist>
       <group sdata="pubDate">
         <count/>
       </group>
       <group sdata="author">
         <count/>
       </group>
    </ctx_result_set_descriptor> ', rs); The output of that query (returned in the value "rs") is:
<ctx_result_set>
  <hitlist>
    <hit>
      <score>3</score><rowid>AAAPoEAABAAAMWsAAC</rowid>
      <sdata name="AUTHOR">John</sdata>
      <sdata name="PUBDATE">2001-01-03 00:00:00</sdata>
    </hit>
    <hit>
      <score>3</score><rowid>AAAPoEAABAAAMWsAAG</rowid>
      <sdata name="AUTHOR">John</sdata>
      <sdata name="PUBDATE">2001-01-03 00:00:00</sdata>
    </hit>
  </hitlist>
  <count>100</count>
  <groups sdata="PUBDATE">
     <group value="2001-01-01 00:00:00"><count>25</count></group>
     <group value="2001-01-02 00:00:00"><count>50</count></group>
     <group value="2001-01-03 00:00:00"><count>25</count></group>
   </groups>
   <groups sdata="AUTHOR">
     <group value="John"><count>50</count></group>
     <group value="Mike"><count>25</count></group>
     <group value="Steve"><count>25</count></group>
   </groups>
</ctx_result_set>