Description
Currently we parse indexing SQL statements with our own homegrown parser. It needs to be updated in order to support the new kinds of indexes. Discussion here shows some of deficiencies. In particular the fn is_same_kind_columns in index.rs has to jump through some hoops in order to compare if the two expressions are "kind of same". For instance BlockRangeLower and Unknown("lower(block_range) int4_minmax_multi_ops)"). Also the test generate_ddl() has a commented out line in the mock_thing_index_list for the same reason.
Related to this is generator of SQL statements that is not perfectly sanitising it's output, somewhat addressed in this.
One approach would be to extend the current parser to handle those cases. Other would be to use sqlparser or similar.
Are you aware of any blockers that must be resolved before implementing this feature? If so, which? Link to any relevant GitHub issues.
No response
Some information to help us out
Description
Currently we parse indexing SQL statements with our own homegrown parser. It needs to be updated in order to support the new kinds of indexes. Discussion here shows some of deficiencies. In particular the
fn is_same_kind_columnsin index.rs has to jump through some hoops in order to compare if the two expressions are "kind of same". For instanceBlockRangeLowerandUnknown("lower(block_range) int4_minmax_multi_ops)"). Also the test generate_ddl() has a commented out line in themock_thing_index_listfor the same reason.Related to this is generator of SQL statements that is not perfectly sanitising it's output, somewhat addressed in this.
One approach would be to extend the current parser to handle those cases. Other would be to use
sqlparseror similar.Are you aware of any blockers that must be resolved before implementing this feature? If so, which? Link to any relevant GitHub issues.
No response
Some information to help us out