BigQuery: Fix column identifier reserved keywords list #1678
Conversation
iffyio
force-pushed
the
bigquery-reserved-columns
branch
from
bff25db to
a4d445b
Compare
src/dialect/mod.rs
Outdated
| /// for the dialect. | ||
| /// For example given: `SELECT <col> AS <alias> FROM T`. The returned | ||
| /// keywords are not allowed in `<col>` and `<alias>` | ||
| fn get_reserved_keywords_for_column_identifier(&self) -> &'static [Keyword] { |
There was a problem hiding this comment.
Theoretically, a dialect may have a different list of reserved keywords for column identifiers and column aliases, and in some cases, the parser will need to look ahead to make that decision accurately.
Perhaps we should adopt a unified approach to this check along the lines of Dialect::is_select_item_alias?
There was a problem hiding this comment.
I agree this would be nice -- perhaps we can do so in a follow on PR
There was a problem hiding this comment.
I think this sounds good! I'll take a look at it tomorrow and I'll ping to take another look at the PR
There was a problem hiding this comment.
I've updated this now to be similar to is_select_item_alias, taking in a parser, PTAL!
| @@ -213,6 +213,15 @@ fn parse_raw_literal() { | |||
| ); | |||
| } | |||
|
|
|||
| #[test] | |||
| fn parse_big_query_non_reserved_column_alias() { | |||
| let sql = r#"SELECT OFFSET, EXPLAIN, ANALYZE, SORT, TOP, VIEW FROM T"#; | |||
src/dialect/mod.rs
Outdated
| /// for the dialect. | ||
| /// For example given: `SELECT <col> AS <alias> FROM T`. The returned | ||
| /// keywords are not allowed in `<col>` and `<alias>` | ||
| fn get_reserved_keywords_for_column_identifier(&self) -> &'static [Keyword] { |
There was a problem hiding this comment.
I agree this would be nice -- perhaps we can do so in a follow on PR
Parser currently uses the same `RESERVED_FOR_COLUMN_ALIAS` keywords list to avoid lookahead when parsing column identifiers. This assumed that all listed keywords were reservered across all dialects which isn't the case. So that the following valid BigQuery statement previously failed due to `OFFSET` being flagged as a keyword. ```sql SELECT 1, OFFSET FROM T ``` This updates the parser to support dialect specific `RESERVED_FOR_COLUMN_ALIAS` list
iffyio
force-pushed
the
bigquery-reserved-columns
branch
from
a4d445b to
2b00afd
Compare
Parser currently uses the same
RESERVED_FOR_COLUMN_ALIASkeywords list to avoid lookahead when parsing column identifiers. This assumed that all listed keywords were reservered across all dialects which isn't the case.So that the following valid BigQuery statement previously failed due to
OFFSETbeing flagged as a keyword.This updates the parser to support dialect specific
RESERVED_FOR_COLUMN_ALIASlist