The labels in the column "reduced" of the Data Set Chooser mark each dataset with "ok / incomplete / -". These labels indicate whether a data set has previously successfully been reduced at least once, all previous reductions failed or a reduction has never been tried. Data sets are identified by their name, which is derived from the first science file within the data set. As long as the data set name is preserved (i.e. the first science file in a data set has not changed), the Data Organizer will consider it to be the same data set. The Data Organizer recognizes any previous reductions of data sets it considers to be the same as the current one, and labels the current data set with "OK" if any of them was successful, even if the previously reduced data set differs from the current one.

Note that the Product Explorer at the end of the reduction will list all the previous reductions of a particular data set. This list might include successful and/or unsuccessful reduction runs with different parameters, or in your case with different input files. The important fact is that these are all reductions of data sets with the same first raw science file. By browsing through all reductions of a particular raw science file, the users can choose the one they want to use.