apache / apache/datafusion-sqlparser-rs
A bracket-quoted identifier containing `]]` round-trips to SQL that fails to reparse
- Dominant language
- Rust
- Stars
- 3.5k
- Forks
- 772
- Avg merge
- 4d 9h
- Merged PRs (30d)
- 17
Description
```rust
use sqlparser::dialect::MsSqlDialect;
use sqlparser::parser::Parser;
fn main() {
let sql = "SELECT [a]]b]";
let ast = Parser::parse_sql(&MsSqlDialect {}, sql).unwrap();
let printed = ast[0].to_string();
println!("input: {}", sql);
println!("printed: {}", printed);
let reparsed = Parser::parse_sql(&MsSqlDialect {}, &printed);
println!("reparsed: {:?}", reparsed);
}
```
```
input: SELECT [a]]b]
printed: SELECT [a]b]
reparsed: Err(ParserError("Expected: end of statement, found: ] at Line: 1, Column: 12"))
```
`[a]]b]` is a bracket-quoted identifier whose value is `a]b` (the tokenizer folds the doubled `]]` into a literal `]`). Displaying the parsed AST back to SQL writes `[a]b]` instead, and that string does not parse as the same identifier: it parses as `a` followed by leftover tokens `]b]`.
Tested on sqlparser 0.62.0.
BTW, this bug was found using [hegel](https://crates.io/crates/hegeltest). Happy to contribute the tests if you're interested.
Contributor guide
No contributing guide indexed for this repository
Research direction
Reproduce the issue with the Rust example using MsSqlDialect, Parser::parse_sql, and AST to_string. Inspect how the tokenizer handles doubled brackets and how the parsed identifier is displayed, then add a round-trip regression test showing that [a]]b] reparses as the identifier a]b and run the relevant test suite.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust
- Domain
- databases
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Quiet
- Clarity
- Clearly specified
- Newbie friendliness
- 68/100