snowflake column identifier for copy into / query data staged in files

According to documentation, Snowflake allows:
* [query data staged in files](https://docs.snowflake.com/en/user-guide/querying-stage.html#required-parameters) with an identifier of a form `[.]$[.]`
* [copy into table](https://docs.snowflake.com/en/user-guide/data-load-transform.html) with an column identifier containing `$` character like `select t.$1,t.$2,t.$3 from @~/datafile.csv.gz t;`

Although the feature seems to be snowflake specific, it cannot be implemented within snowflake dialect. It's because tokens starting with `$` character are tokenized into `Token::Placeholder`. That's why parser's change is required.

Initial OpenLineage project issue: https://github.com/OpenLineage/OpenLineage/issues/814

lexer と parser から始め、`$` で始まるトークンの処理、特に現在 `Token::Placeholder` を生成している経路を確認し、Snowflake 方言の処理を見直します。文書化されている staged-file と COPY INTO のカラム形式が既存のプレースホルダー構文を後退させることなく正しくパースできれば、作業は完了です。

索引モデルが issue の本文から書いたものです。