We need a fully scriptable import script step.
Which means that every_thing we can set would accept a calc field.
The matching field would defined by a list of values
fieldIDs : a value list with the fiedlIDS in the correct matching order, a blank value = skip, an "*" before the fieldID would denote is as matching field
And we could specify if we want to ignore indexing altogether, or have it done after the whole importing (so it would be faster), or like it is today
this would allow countless of usage, including having the user set import matching in a nice filemaker interface made by the dev, so the user would be able to do the field matching in a friendly manner