hreddygayam opened a new pull request, #18584:
URL: https://github.com/apache/iceberg/pull/18584
## What
`TableMetadata.newTableMetadata(...)` always reassigns fresh, sequential
field IDs via `TypeUtil.assignFreshIds`, so there is currently no supported way
to create table metadata that keeps the field IDs already set on the provided
schema. This matters when the IDs are already meaningful — for example
migrating or re-registering a table whose data files were written against a
specific ID assignment.
This adds a public overload with a `reassignFieldIds` flag:
```java
public static TableMetadata newTableMetadata(
Schema schema, PartitionSpec spec, SortOrder sortOrder,
String location, Map<String, String> properties, boolean
reassignFieldIds)
```
- `reassignFieldIds = true` — what every existing overload now delegates to
— keeps today's behavior exactly.
- `reassignFieldIds = false` preserves the schema's field IDs: the schema is
rebuilt with the initial schema ID, `last-column-id` is taken from
`schema.highestFieldId()` so later column additions don't collide, and the IDs
are validated to be positive (uniqueness is already guaranteed by `Schema`
construction). The partition spec and sort order are still rebuilt against the
resulting schema, so their source IDs are preserved as well.
## Why
Closes #17614.
As discussed on the issue, this is intentionally a narrow, expert-facing
method rather than a `Catalog.TableBuilder` option — it avoids a flag that
other catalog implementations (REST client, engine catalogs) would silently
ignore, and avoids a REST OpenAPI `CreateTableRequest` change. A `TableBuilder`
option can follow in a separate PR if there's appetite.
cc @RussellSpitzer — this is the "new method for experts" shape you
suggested on the issue; does the signature look right before I invest further?
## Tests
Added to `TestTableMetadata`:
- preserves field IDs (including nested fields and a partition source
column),
- `reassignFieldIds = true` matches the existing fresh, sequential ID
assignment,
- non-positive field IDs are rejected.
`:iceberg-core:test`, `:iceberg-core:checkstyleMain`, and
`:iceberg-core:revapi` all pass (the new public static method is a source- and
binary-compatible addition).
--
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]
