Version 3.0 of SQLite will support unicode in three different
representations: UTF-8, UTF-16LE, and UTF-16BE. So the question
has come up about what encoding(s) should the COPY command support.
The COPY command, you will recall, is a non-standard command in
SQLite that is modeled after the same command in PostgreSQL. The
SQLite version is more limited than the PostgreSQL version. COPY
in SQLite allows you to quickly copy import amounts of data from
a text file into an internal table. The original PostgreSQL
version allows import and export.
Here are the options:
(1) Extend the syntax of the COPY command to allow the user
to specify the text representation of the file being
imported.
(2) Always assume the import is coming from a UTF-8 file.
(3) Remove the COPY command all together.
Right now, we are leaning toward option (3) - delete the COPY
command. If you think that is the wrong approach, this is your
opportunity to speak up.
Reasons to remove the COPY command:
* It is not standard SQL.
* You can accomplish the same thing just as quickly using BEGIN, a
precompiled INSERT statement executing multiple times with
different bindings each time, and a COMMIT.
* Removing COPY makes SQLite simpler and smaller and we very
must want to keep SQLite simple and small.
* The equivalent functionality can be moved into the shell.
Instead of a COPY command in SQL, you can have a ".copy"
shell command. And the new ".copy" shell command (actually,
it will probably have a different name) can support formats
other than the PostgreSQL format.
--
D. Richard Hipp -- [EMAIL PROTECTED] -- 704.948.4565
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
