lunar-shadow opened a new issue, #14386:
URL: https://github.com/apache/iceberg/issues/14386
### Feature Request / Improvement
### Description
When using the **Iceberg REST catalog** with services like **Polaris**,
schema and table names are stored **preserving their exact case** (e.g.,
`CustomerDB.Orders`).
As a result, queries that use different casing (e.g., `customerdb.orders`)
fail with `SCHEMA_NOT_FOUND` or `TABLE_NOT_FOUND`, even though the underlying
object exists.
This behavior causes friction for users and tools that expect
case-insensitive resolution (as seen in other engines like Trino, Spark, and
Snowflake).
---
### Steps to Reproduce
1. Configure Spark to use a REST catalog pointing to Polaris.
2. Run:
```sql
CREATE TABLE test_catalog.CustomerDB.Orders (id INT);
SELECT * FROM test_catalog.customerdb.orders;
```
3. Query fails because identifiers are matched case-sensitively.
---
### Expected Behavior
Iceberg should optionally allow **case-insensitive identifier matching** for
namespace, table, and view lookups when interacting with REST catalogs that
preserve case.
---
### Proposed Solution
Introduce an optional configuration flag (disabled by default):
```
iceberg.rest-catalog.case-insensitive-name-matching=true
```
When enabled, lookups in the REST catalog (namespace, table, view) should be
performed case-insensitively.
---
### Additional Context
* Polaris stores identifiers with case preserved.
* Gravitino has a similar enhancement request:
[https://github.com/apache/gravitino/issues/1853](https://github.com/apache/gravitino/issues/1853)
* Trino implemented case-insensitive matching at the connector layer:
*
[https://github.com/trinodb/trino/issues/23715](https://github.com/trinodb/trino/issues/23715)
*
[https://github.com/trinodb/trino/pull/23867](https://github.com/trinodb/trino/pull/23867)
This issue requests adding the logic inside the **Iceberg REST catalog
layer**, so clients don’t need to handle it individually.
---
### Environment
* Spark: 3.5.x
* Iceberg version: (current master / 1.x)
* Catalog: REST (Polaris)
* Behavior observed on both driver and executor when case differs.
---
### Impact
* Affects interoperability between Iceberg and Polaris.
* Users must currently match exact case to query objects.
* Impacts mixed-client ecosystems (Spark, Flink) that normalize identifiers
differently.
### Query engine
Spark
### Willingness to contribute
- [x] I can contribute this improvement/feature independently
- [ ] I would be willing to contribute this improvement/feature with
guidance from the Iceberg community
- [ ] I cannot contribute this improvement/feature at this time
--
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]
