JeonDaehong opened a new issue, #3381:
URL: https://github.com/apache/iceberg-rust/issues/3381
### Apache Iceberg Rust version
v0.10.0
### Describe the bug
`PartitionSpec::partition_to_path` panics when a partition field's source
column is no longer in the schema and the field is `identity` or `truncate`
over a non-string type.
#2773 made `partition_type` fall back to a source-independent result type
for dropped source columns (`Unknown` for identity/truncate), so the
`partition_type(..).unwrap()` succeeds. The panic happens later:
`to_human_string` falls back to `Display for Datum`, which has no arm for an
`Unknown`-typed int/long and hits `unreachable!()` (`datum.rs`).
Java's `PartitionSpec.partitionToPath` handles the same case: `resultType`
uses `UnknownType` for a dropped source and `toHumanString` falls back to
`value.toString()`.
| Old spec field, source dropped | Rust | Java |
| --- | --- | --- |
| identity(int) | panic | `p=1` |
| identity(timestamp) | panic | `p=1` |
| truncate(int) | panic | `p=10` |
| identity(string), bucket, truncate(string), day, void | ok | ok |
Today this is reachable through the public `partition_to_path`. #3253 also
starts summarizing deleted files under their historical spec, which would put
it on the commit path (see the discussion there).
### To Reproduce
```rust
use iceberg::spec::{Literal, NestedField, PartitionSpec, PrimitiveType,
Schema, Struct, Transform, Type};
let old_schema = Schema::builder()
.with_fields(vec![
NestedField::required(1, "x",
Type::Primitive(PrimitiveType::Int)).into(),
NestedField::required(2, "y",
Type::Primitive(PrimitiveType::Int)).into(),
])
.build()
.unwrap();
let spec = PartitionSpec::builder(old_schema)
.add_partition_field("y", "p", Transform::Identity)
.unwrap()
.build()
.unwrap();
// `y` has been dropped from the current schema.
let current = Schema::builder()
.with_fields(vec![NestedField::required(1, "x",
Type::Primitive(PrimitiveType::Int)).into()])
.build()
.unwrap();
// Panics: internal error: entered unreachable code (datum.rs)
spec.partition_to_path(&Struct::from_iter([Some(Literal::int(1))]),
current.into());
```
### Expected behavior
No panic. Format the value as Java does for an unknown source type (`p=1`,
`p=10`).
### Willingness to contribute
I can contribute a fix for this bug independently
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
