OK, what I would do (and assuming you only need to do this once) is . . . .

Assuming your data is ALREADY sorted by column I, and so that the last row for each set of duplicates is the row you want.

Create a copy of your data, just in case this goes wrong, so that you don't lose your data.

Apply a helper column to identify the last row for each set of duplicates, like this, assuming your data starts in cell I2.

In some other column, say L . . .

=I2=I3

Copy this formula down as far as required.

You should now have a FALSE value in column L, for the LAST item for each set of duplicates (including sets containing only a single item), and a TRUE value for every other item.

Select all the formulas in column L, and use Copy, Past Special, Values, to convert the formulas to values.

Then select ALL your data (columns A to L), and use Data, Sort, using column L as the primary sort key.

This should bring all the TRUE / FALSE values together, and it should be simple to delete all the TRUE values.

Then if necessary, re-sort using column I as the primary sort key.