[Freebase-discuss] Query on interpreting quad dump.

[sharath jagannath]

Hey,

I am trying to build a data ingestor and using quad dumps to achieve it. I
have problem figuring our couple of things from the dumps and does not want
to make my ingestor sub-optimal.

I will describe the problem with an example (actual tuple from the dump):

*/m/01l25yw      /type/object/name       /lang/en        Academia de danças*
*
*
This is an instance of the musical album and I want to map it to a RDF
literal.

/type/object/name's
<http://www.freebase.com/type/schema/type/object>expected type is
text <http://www.freebase.com/type/schema/type/text>, which has
language<http://www.freebase.com/type/schema/type/lang>as its
property, finally language has iso639 as its property and
expects enumeration.
As you can see the dump gives "/lang/en" for "to" and "Academia de danças"
for "value".

I do not know how to map this chain of
properties (name->language->iso639) to a single property.
As of now I am considering the "/lang/en" (has a special case) directly to
be language to be used for creating literal (RDF world), which works fine,
but this is more of a code driven approach which I have now and want it to
be configuration driven.

For this particular example language has only one property iso639, but say
key <http://www.freebase.com/type/schema/type/key>->namespace<http://www.freebase.com/type/schema/type/namespace>
contains
multiple property and I did not know how to interpret:
*/m/01k975n      /type/object/key        /authority/musicbrainz
 2a39c54a-bd0d-3299-b60a-37196650c963*
*
*
I understand that "/authority/musicbrainz" is again an enumeration but, I am
not sure which other properties are to be considered as a special case.
Would appreciate if anybody suggests a better configuration driven approach.
*
*

*
*

Thanks,
Sharath
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freebase.com/pipermail/freebase-discuss/attachments/20101019/f2d054be/attachment.htm>