Populate custom field values

Sep 2, 2008 at 5:10 PM

I would like to populate custom field values using the following methods

1) Populate during content feeding time as document attributes

Other than default document field values like title, url and body i would like to have my own custom fields. I want to fill that values using filetraverser.

How do i do that? 

2) Populate from existing fields

In my sample document i may have several cars like Accord, Camry, Acura, Lexus. In the car entity field i want to extract the first car name that is mentioned first in the document.

How do i do that?

Thanks for your help in advance.

Sep 3, 2008 at 5:21 AM

Question 1) Populate during content feeding time as document attributes
Answer:
    1. Create the new custom fields in IndexProfile.xml file of FAST
    2. See the pipeline that you are using to feed the documents. Add a custom stage to this pipeline and configure it such that when you use filetraverser, it should be able to populate the custom fields.
    3. Run FileTraverser to feed the documents.
Note: Step 2 is a cold update and all the existing Index' maintained for the documents already fed in FAST will be lost. So, you will have to re-feed all the documents.

Question  2) In my sample document i may have several cars like Accord, Camry, Acura, Lexus. In the car entity field i want to extract the first car name that is mentioned first in the document.?
    Option 1: If you use the default entity extraction stage, it will get you all the cars from the document and put into the cars field.
    Option 2: Create a new stage (using python), that should check the body of each document being fed against the cars master (database/xml). Compare and get the first occurance and write it back to the in-memory 
    document object.
Option 2 will take care of your precise requirement, but it will be more time consuming that option 1.

So, basically all the work to be done is in FAST and nothing outside it.

Hope this helps!!!
Sep 3, 2008 at 1:21 PM
Edited Sep 4, 2008 at 4:29 PM
Thanks for your help

I think i didn't clearly mention in my 1st question

In my 1st question i want to fill the custom field out of my document content. While feeding time i like to fill dynamic attributes like author name company name that is not present in the document

Thanks again
Jul 20, 2009 at 2:44 PM

a little forum-bump here:)

I managed to do a cold update and all my collections are now empty.

How do i refeed them? I have all the document data, but the documents are not in the index. When the crawler finds the document, it thinks it is unchanged and it doesn't index it again.

I'll appreciate any help. Thanks!