Get orphaned extension rows

[pieterprovoost]

I need to do some validation on the archives I'm reading, and this includes checking if all extension rows have matching core rows. This PR adds a method DwCAReader.orphaned_extension_rows() which returns non-matching core IDs and orphaned extension row indices as follows:

{
    'extendedmeasurementorfact.txt': {
          u'Cruise68:Station593:EventSorbeSledge9887:Subsample16686_5': [11136],
          u'Cruise64:Station550:EventSorbeSledge9740:Subsample9878_4': [1413, 1414, 1415]
    },
    'occurrence.txt': {
        u'Cruise66:Station591:EventSorbeSledge9885:Subsample9959_2': [4928],
        u'Cruise66:Station589:EventSorbeSledge9883:Subsample9953_1': [4739],
        u'Cruise68:Station579:EventSorbeSledge9810:Subsample17412_3': [5977]
    }
}

There may be a more elegant implementation.

[niconoe]

Thanks a lot @pieterprovoost, glad that my work is useful!

I'm considering merging this PR, but I'll need at least some documentation (API, a "performance" warning stating that this method can be time and memory consuming, ...), some unit tests and an entry in the changelog.

Would you be interested in working on that? Otherwise, I'll give it a try but I can't promise when ;)

Thanks again

[pieterprovoost]

I'll see what I can do, may take a while as well :)

[niconoe]

Manually merged, thanks a lot!