The first record is quite different to the following records - this meant you might need several load-steps to the header and the records and to match them in any way. Here a suggestion how the following records could be loaded:
LOAD subfield([@1:n], ':', 1) as F1, subfield([@1:n], ':', 2) as F2, subfield([@1:n], ':', 3) as F3
FROM [test.txt] (fix, codepage is 1252, header is 1 lines);
In a similar way (with subfield() and/or other string-functions) would it be possible to load also the first record.