If you are updating an index column on 1 million rows, that is perfectly legal, but there may be performance implications.
For each update, the existing index entry would be removed and a new one created at a different location in the index.
There may also be an issue regarding clustering of the table, but I am unsure if it applies to updates (but does apply to inserts). If the index being updated was defined as the clustering index, then DB2 "might" move the row in the table to the new location to maintain the clustering sequence. But as I said, I am not sure if clustering applies to updates like it does apply to inserts.
In any case, you might want to define a large amount of freespace for the index and/or table to handle the large update, and perform a reorg after the update is finished.