fix(connector-iceberg): partition logic hardcodes dummy values and ignores table spec

[ashudeqode]

Bug description

Description of Bug
The Apache Iceberg sink routing logic (write_data in router/mod.rs) contains severe bugs in partition key construction:

Hardcoded Zero/Dummy Partition Values: primitive_type_to_literal returns static zero/empty values (PrimitiveLiteral::Int(0), PrimitiveLiteral::String(String::new()), etc.). When passed into get_partition_type_value and PartitionKey::new(...), every batch is written under a single epoch-zero/dummy partition regardless of payload contents.

Empty PartitionSpec Construction: The code calls PartitionSpec::builder(table.current_schema_ref()).build(), which creates an unpartitioned spec (Spec ID 0) rather than retrieving the active spec via table.metadata().default_partition_spec().

Missing Record Fanout: DataFileWriter expects a single PartitionKey per writer instance. There is currently no logic to evaluate partition transforms against incoming RecordBatch data or group records across target partition writers.

Area
Connectors (iggy-connector-sdk / Iceberg connector) — High-Risk Area

Steps to Reproduce
Point the Iceberg sink to an Iceberg table partitioned by a field (e.g., identity(date) or day(event_timestamp)).

Produce JSON messages with varying timestamp/date payload fields into Iggy.

Observe sink output: all records end up in epoch 0 / default fallback partition paths.

Affected area / component

Connectors

Deployment

Docker (self-built image)

Versions

No response

Hardware / environment

No response

Sample code

No response

Logs

No response

Iggy server config

No response

Reproduction

No response

Contribution

• I'm willing to submit a pull request to fix this bug

Good first issue

• I think this could be a good first issue for a new contributor