Because the way MySQL does other things seems has been troublesome. IIRC, MySQL essentially generates incremented integer table and column ids, so rebuilding tables relationships meant that you had to manually regenerate the table schema, then update all the old table files with the new table and column ids which were incremented on from what they were previously.

With Postgres, you can just drop the binary table data files in, restart the database and pg rebuilds the indexes and relationships.

From a quick google, it now looks like MySQL is now more robust in this regard.