Querying metadata

[aleks]

Hello,

 

Does anyone know how best to construction a query url to find results that contain a specific character within a metadata field?

 

For example:

 

I have some items with meta_N set as; MATXGTL106M, MATXGTL107M, MATXGWL108M

 

How do I query against meta_N to only retrieve the items with TL in them?

 

I've tried truncation operators with no luck.

[Benjamin_Pearson]

You need to pass the meta into the query eg. meta_N_trunc=*TL*

 

See http://docs.funnelback.com/metadata_search_field.html

[aleks]

Hi Ben,

 

Thanks! however, it's still not quite right.

 

This is part of my query...

 

&fmo=false&num_ranks=100&form=json&meta_N_trunc=*TL*

 

'No results returned' :(

[gordongrace]

Hi Aleks -

 

I suspect that you're making the correct request there - your syntax appears correct.

 

Looking closely at the metadata query operator documentation page, you may see:

 

Please note - Word truncation is supported only when the -service_volume=low query processor option is set

 

 

There's a related query processor option (service volume) that needs adjustment before truncated queries can be used:

 

- http://docs.funnelback.com/13.2/service_volume_settings.html

- http://docs.funnelback.com/13.2/query_processor_options_collection_cfg.html#G.%20Query%20interpretation%20options

 

Truncated queries are typically much more computationally expensive to process - they're generally suited only to low query volume collections (or where users are happy to wait a little longer for their results to be returned).  This impact on performance also necessitates the QPO not being controllable as a CGI parameter.

[Suhu]

Hello guys,

 

is it possible to get the meta data information for a page, if we know the URL of the page ?

 

Thanks

[gordongrace]

From the command line, $SEARCH_HOME\bin\padre-di will address that issue for you.

http://docs.funnelback.com/13.2/padre_binaries_usage.html#23.%20padre-di

 

Using the Public UI, your best approach may be to conduct a single-result query using the domain and path portions of the URL to constrain the query:

 

http://search-demo-au.funnelback.com/s/search.json?collection=demo-v13-recommender&meta_u_sand=www.bookdepository.com&meta_v_phrase_sand=Processing-Casey-Reas%2F9780262182621&num_ranks=1&sort=url&SM=meta&SF=abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLNMNOPQRSTUVWXYZ1234567890

 

For a web collection, the default metadata mapping uses 'u' for the domain portion of the URL and 'v' for the path portion of the URL.  Sorting the results by URL and limiting the number of results to one will ensure we don't include child pages in the response.  I've also set the summary mode to metadata, displaying all available fields

[gordongrace]