This project demonstrates schema evolution issues in Apache DataFusion when reading multiple files with incompatible schemas. It compares the behavior between standard Parquet format and Vortex format.
cargo r --example parquet+----+------+-------+
| id | code | value |
+----+------+-------+
| 1 | A100 | 100 |
| 2 | B200 | 200 |
| 3 | C300 | 300 |
| 4 | 400 | 400 |
| 5 | 500 | 500 |
| 6 | 600 | 600 |
+----+------+-------+
Query succeeded unexpectedly
cargo r --example vortexthread 'tokio-runtime-worker' (110661964) panicked at /Users/xxx/.cargo/git/checkouts/vortex-0eff073da78b9952/d9fffbe/vortex-error/src/lib.rs:310:33:
Failed to convert scalar to utf8:
Expected a string scalar, found Primitive(I64(400))
Backtrace: