apache / apache/datafusion-sqlparser-rs
Expose a method for mutating Parser::index
- Dominant language
- Rust
- Stars
- 3.5k
- Forks
- 772
- Avg merge
- 4d 9h
- Merged PRs (30d)
- 17
Description
We have designed an SQL parser using sqlparser's `Parser`, and we wish to implement two helper methods for consuming tokens—similar to `Parser::consume_token` and `Parser::consume_tokens`—but with the ability to accept string inputs and perform case-insensitive matching.
However, since the Parser does not provide a method to mutate the `Parser::index`, we are unable to correctly implement our own `consume_tokens` because we need to revert the parser state if consumption fails mid-way.
We propose adding and exposing a new method in sqlparser's `Parser`, such as `set_index` or `index_mut`, to allow users to mutate the index.
The expected implementation of our `consume_token` and `consume_tokens`:
``` rust
/// Consumes the next token if it matches the expected token, otherwise return false.
///
/// Note, the matching is not case sensitive.
fn consume_token(&mut self, expected: &str) -> bool {
if self.parser.peek_token().to_string().to_uppercase() == *expected.to_uppercase() {
self.parser.next_token();
true
} else {
false
}
}
/// If the current and subsequent tokens exactly match the `tokens` sequence, consume them and returns true.
/// Otherwise, no tokens are consumed and returns false
///
/// Note, the matching is not case sensitive.
fn consume_tokens(&mut self, tokens: &[&str]) -> bool {
let index = self.parser.index();
for token in tokens {
if !self.consume_token(*token) {
// Or: `*self.parser.index_mut() = index;`
self.parser.set_index(index);
return false;
}
}
true
}
```
Contributor guide
No contributing guide indexed for this repository
Research direction
Start by locating the Rust Parser implementation and its existing index, consume_token, and consume_tokens methods. Check the parser tests around token consumption, then expose a way to mutate or set the index so failed multi-token consumption can restore the prior position.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust
- Domain
- databases
- Issue type
- Feature
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100