Hi,
I'm trying to get Funnelback to index data as CSV and cannot get it to make sense of the content.
Here is my current config:
The content is in Matrix, text formatted as CSV, at a normal url (headers set to text/html; charset=utf-8). My CSV looks like:
email@email.com,Mr,first name,last name,1234,123,S234,job title,,,,department
FB is looking at a a standard page in Matrix (blank design and paint layout) of
I've tried using a text file in Matrix, as opposed to a standard page, so it gets the correct header (text/csv) but couldn't get FB to actually index that file at all.
My collection.cfg has a few options set there, including:
indexer_options=-csv=,n -csv_fields=T1,A1,F1,L1,N1,A1,R1,J1,-X,-X,-X,D1
I have two main issues:
1. getting FB to create multiple documents in that collection if i have multiple lines in my CSV. That doesn't work at the moment. If i have more than one line, it will simply create one record on index and have the content from both CSV lines
2. getting FB to make sense of my metadata mapping. At most I could get it to create one metadata field for the record it creates, (as "t"). I'm not sure how it managed that. In most cases it simply doesn't translate the CSV columns to metadata classes.
I'm using a web collection. My metamap.cfg and xml.cfg files are empty. And the collection besides this is pretty standard. It was initially set up to index the data as XML and was working, so I've just tweaked it to grab the data as CSV.
Is indexing CSV with FB even possible? There's very little documentation about this, so I'm not entirely sure it's been used or tested, or is even a feature.
Gilles
