Index attachments in MSG files

Question

I am trying to use Oracle Text to index and search msg files with attachments. For testing purposes, I have set up a table called TEST_DOCUMENTS. The table has two columns - a blob column containing the data to be indexed and a varchar2 column describing the data as follows:
Description - Test 1 - simple email no attachments
Blob - Outlook MSG file.
Description - Test 2 - an email including a PDF attachment
Blob - Outlook MSG file.
Description – Test 3 - an email including a Word attachment
Blob - Outlook MSG file.
Description – Test 4- a PDF document
Blob – Adobe PDF file.
Description – Test 5 - a Word document
Blob – Ms Word Doc file.
When I index the above using the following:
CREATE INDEX TEST_DOCUMENTS_IDX ON TEST_DOCUMENTS
(DATA_BLOB )
INDEXTYPE IS CTXSYS.CONTEXT PARAMETERS ('filter ctxsys.auto_filter');
And search using
Select description
from test_documents
where contains(data_blob,:’search string here’) > 0;
The email attachments are not included in the search, i.e., I can search the content of the emails but not the attachments. Note: Test 4 and Test 5 are the same as the attachments linked to Test 2 and Test 3. These are returned when the appropriate string is entered!
If I drop the index and create it as:
CREATE INDEX TEST_DOCUMENTS_IDX ON TEST_DOCUMENTS
(DATA_BLOB )
INDEXTYPE IS CTXSYS.CONTEXT PARAMETERS ('filter ctxsys.mail_filter');
I can search the email with the word attachment but not the body of the email. Test 2 with PDF attachment is not indexed. Test 4 and Test 5 as expected are not indexed as they are not email documents.
On checking the drmailfl.txt file I can see that the default behavior is as follows:
[behavior]
text          INCLUDE
application     IGNORE
application/msword AUTO_FILTER
image          IGNORE
audio          IGNORE
video          IGNORE
model          IGNORE
i.e., only process word attachments. However, if I add application/pdf to the file it makes no difference. The PDF attachment is not indexed.[ Note: each time I make a change I drop and recreate the index..] I even tried setting the behavior to ignore for everything! You would expect it to not index anything but it still indexed the word attachment! I tried shutting down the database and restarting it to see if it made any difference.
The documentation for Oracle 11g does imply that the ‘AUTO_FILTER’ should index the attachments, but I cannot get it to do it. Has anybody successfully indexed emails with PDF or Word attachments using either ‘AUTO_FILTER’ or ‘MAIL_FILTER’?

Answer 1

Hi James
Strictly speaking, you can't. The Workflow system has its own mechanism for handling attachments - they are attached via Form Manager's custom "attachment" user interface, rather than being attached "inside" the PDF.
You could get at these attachments, and attach them to your outbound email as regular attachments, rather than embedded within the PDF.
If you have a strong requirement to embed the attachments inside the PDF, there may be something we could do. Please email info-at-avoka.com if you want to discuss further.
Howard
http://www.avoka.com