Input Option 1 - Country centric
In its simplest mode you have a field with country information only, that could be a country code like "US", "DE" or a country name like "Germany". Spelling does not matter, it can be different in each input row, everything...
mapped to LAND1 in this example
Performance: see CountryID - Country centric
Input Option 2 - City centric
The minimum input is
LOCALITY1 with City information
In this mode the transform tries to guess the country that city belongs to. Obviously this mode is not very reliable and works for large cities only.
The more information the transform gets to more often it can identify the country like by adding
COUNTRY as name or code
REGION1 with state or territory
it can actually do due dilligence checks like Vienna, US, 1020 does not match. The postal code points towards Vienna in Austria, the country however is US and there is a Vienna in US also - the Confidence Score is 77 for US.
Or in what country is "Viena" with postal code "1020"? That is Austria with score 100 although the city name was mistyped.
Performance: see CountryID - City centric
Input Option 3 - unstructured in one line
In case above information is not available in distinct fields but all in one line like "1020 Vienna, Austria" all is put into the
Performance: see CountryID - all in one line
Input Option 4 - unstructured in multiple lines
And finally, you might had a source with AddressField1 - 12 but no idea where the country is, what field the city contains or maybe it is different for each row?
is meant for that.
Performance: see CountryID - Multilines
As data this transforms generates regardless of its input is
There are no real options to specify. It is using the Address directory so requires a Directory Path to that, but by default this field is set to the substitution variable we have set correctly during the installation. And "Run as Separate Process" is not required as this transform does nothing in memory.