For the Classify service at OCLC, when it is LCC we use a regular
expression: "^[a-zA-Z]{1,3}[1-9].*$". For DDC we filter out the
truncation symbols, spaces, quotes, etc.

-Steve

On Wed, Jun 20, 2012 at 8:54 AM, Arash.Joorabchi <[log in to unmask]> wrote:
> Hi all,
>
> I am using the OCLC Classify API. As show in the sample response snippet
> below the two attributes "sfa" and "nsfa" could hold different values.
>
> According to
> "http://oclc.org/developer/documentation/classify/response-details":
>
> sfa - classification number from the subfield $a of 082/092 or 050/090,
> or 060/096
>
> nsfa - normalized classification number from the subfield $a of 082/092
> or 050/090, or 060/096
>
> However,I would like to know how this normalization is done.
>
> Thanks,
> Arash
>
>
> <recommendations>
>
> <graph>http://chart.apis.google.com/chart?cht=p&amp;chs=350x200&amp;chd=
> t:100.0&amp;chtt=All+Editions&amp;chdl=Classified (100.00%)</graph>
>    <fast>
>
> <graph>http://chart.apis.google.com/chart?cht=p&amp;chs=475x175&amp;chd=
> t:100.0,16.666668,16.666668,16.666668&amp;chl=Functional programming
> (Computer science)|Lambda calculus|Modality (Logic)|Type theory|</graph>
>      <headings>
>        <heading heldby="6" ident="fst00936086">Functional programming
> (Computer science)</heading>
>        <heading heldby="1" ident="fst00991011">Lambda
> calculus</heading>
>        <heading heldby="1" ident="fst01024350">Modality
> (Logic)</heading>
>        <heading heldby="1" ident="fst01159972">Type theory</heading>
>      </headings>
>    </fast>
>    <ddc>
>      <mostPopular holdings="6" nsfa="510.78" sfa="510.7808"/>
>      <mostRecent holdings="6" sfa="510.7808"/>
>
> <graph>http://chart.apis.google.com/chart?cht=p&amp;chs=350x200&amp;chd=
> t:100.0&amp;chtt=DDC&amp;chdl=510.7808</graph>
>    </ddc>
>  </recommendations>