In a company, several different programs insert rows into a messages table which should later be parsed and inserted into appropriate tables. Consider the tables:

TABLE messages
  id INTEGER NOT NULL PRIMARY KEY
  type VARCHAR(30) NOT NULL
  data VARCHAR(100) NOT NULL

TABLE notifications
  id INTEGER NOT NULL PRIMARY KEY
  message VARCHAR(100) NOT NULL

TABLE alerts
  id INTEGER NOT NULL PRIMARY KEY
  shortMessage VARCHAR(15) NOT NULL

Create a procedure called ParseMessages, in the default schema, which should do the following for all rows in the messages table:

Insert all rows where the type is 'notification' to the notifications table with the message field being set to the data field of messages table.
Insert all rows where the type is 'alert' to the alerts table with the shortMessage field being set to the first 15 characters of the data field of messages table.

In both cases the id field should be the same as in the messages field.

See the example case here for more details.