Add support for mixed BigQuery table name quoting #971
Conversation
| .flat_map(|ident| { | ||
| ident | ||
| .value | ||
| .split('.') |
There was a problem hiding this comment.
does BigQuery support any way to escape . (like 'foo\.bar' I wonder 🤔 )
There was a problem hiding this comment.
BigQuery didn't support escaping dots in identifiers either as far as I could tell and tested
| @@ -84,9 +86,16 @@ fn parse_raw_literal() { | |||
|
|
|||
| #[test] | |||
| fn parse_table_identifiers() { | |||
| fn test_table_ident(ident: &str, expected: Vec<Ident>) { | |||
| fn test_table_ident(ident: &str, canonical: Option<&str>, expected: Vec<Ident>) { | |||
There was a problem hiding this comment.
can you please document what canonical represents in this context?
|
I wonder if you could respond to @MartinNowak 's comment #889 (comment) as well? Aka does this change really need to be in sqlparser-rs? |
I'm not entirely sure I follow, the changes made in this MR were based off my understanding of what was being proposed in that comment, unless I misunderstood - maybe @MartinNowak can clarify otherwise It wouldn't seem reasonable to do this outside the parser - without it, all users of the BigQuery dialect essentially get back wrong object names from the parser in such cases and will need to know to check all identifiers they get back and fix it up with code similar to the MR - effectively re-parsing the object name themself. |
As described in the gh issue, BigQuery accepts arbitrary quoting identifiers within table names - such that e.g: `'a.b.c'` == `'a'.'b'.'c'` == `'a.b'.'c'` are all equivalent. This adds support for such syntax Fixes apache#889
iffyio
force-pushed
the
bq-quoted-identifiers
branch
from
3ff9d91 to
1084ff5
Compare
|
Thanks! :) |
Co-authored-by: ifeanyi <[email protected]> # Conflicts: # tests/sqlparser_bigquery.rs
Co-authored-by: ifeanyi <[email protected]>
As described in the gh issue, BigQuery accepts arbitrary quoting identifiers within table names - such that e.g:
'a.b.c'=='a'.'b'.'c'=='a.b'.'c'are all equivalent.This adds support for such syntax
Fixes #889