For things like seeding a database, I recommend Richard Brown’s answer: you’re arguably better off using something like Rails seeds mechanism, or something scripted like a rake task.
That said, being able to pipe sql (raw, or a file) is a useful feature, especially for idempotent things like simple look ups or routine queries. In which case you can execute your local sql with any of the following:
$ cat file.sql | heroku pg:psql --app app_name
$ echo "select * from table;" | heroku pg:psql --app app_name
$ heroku pg:psql --app app_name < file.sql