Import data … → same implementation as google_sql.sh?

Hello,

I need to import a bulk of data from a SQL file. 

The data is rather much, around 1 GB in SQL code.

The best method which is least error-prone and best can handle big files
is the google_sql.sh client on Linux, since it seems to fully support UTF-8.

However, it seems to me that running the SQL script from my computer
slows down the whole process because of the roundtrip time for each line to be imported.

I would fancy using the "Import data …" feature from the API console.

But the implementation is different :-( and error messages are not very helpful.

Will there be improvement and consolidation?