chore: introduce new with capacity to align generic row with java side

[luoyuxia]

Purpose

Linked issue: close #186

Brief change log

GenericRow::new() -> GenericRow::new(field_count: usize)

Tests

API and Format

Documentation

[luoyuxia]

cc @leekeiabstraction @fresh-borzoni

[Copilot]

Pull request overview

Updates GenericRow construction to require an explicit field count (pre-initialized to nulls) to better match Java behavior and allow partial rows (e.g., PK-only deletes).

Changes:

• Change GenericRow::new() to GenericRow::new(field_count: usize) and update set_field to overwrite instead of insert.
• Update call sites across core, bindings, examples, and integration tests to pass explicit field counts.
• Modernize a number of error/diagnostic messages to use Rust’s named-format interpolation.

Reviewed changes

Copilot reviewed 18 out of 18 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
crates/fluss/src/row/mod.rs	Introduces GenericRow::new(field_count) with null initialization; changes set_field semantics; updates tests.
crates/fluss/src/row/compacted/compacted_row_reader.rs	Adjusts row creation to pre-sized rows for deserialization.
crates/fluss/tests/integration/kv_table.rs	Updates integration tests to construct rows with explicit field counts; enables PK-only delete rows.
crates/fluss/tests/integration/table_remote_scan.rs	Updates test row construction for append/scan scenario.
crates/fluss/src/record/arrow.rs	Updates unit tests to construct rows with explicit field counts.
crates/fluss/src/client/table/log_fetch_buffer.rs	Updates test row construction with explicit field counts.
crates/examples/src/example_table.rs	Updates example to construct rows with explicit field counts.
crates/examples/src/example_kv_table.rs	Updates KV example to use explicit field counts and demonstrate PK-only delete.
bindings/cpp/src/types.rs	Updates FFI conversion to allocate rows with correct field count.
bindings/python/src/table.rs	Refactors error strings (named interpolation) in Python binding conversions.
crates/fluss/src/row/datum.rs	Refactors error strings (named interpolation).
crates/fluss/src/row/decimal.rs	Refactors error strings (named interpolation).
crates/fluss/src/row/column.rs	Refactors panic message formatting.
crates/fluss/src/metadata/table.rs	Refactors error string formatting.
crates/fluss/src/metadata/partition.rs	Refactors error string formatting.
crates/fluss/src/metadata/json_serde.rs	Refactors JSON serde error string formatting.
crates/fluss/src/client/table/upsert.rs	Refactors validation error string formatting.
crates/fluss/src/client/table/remote_log.rs	Minor formatting / unused-variable tweak and message formatting updates.

Comments suppressed due to low confidence (1)

crates/fluss/src/row/compacted/compacted_row_reader.rs:59

• GenericRow::new(self.row_type.fields().len()) already initializes every field to Datum::Null, so in the nullable/null branch you can skip row.set_field(col_pos, Datum::Null) and just continue;. This avoids an extra bounds-checked write per null field in a hot deserialization loop.

        let mut row = GenericRow::new(self.row_type.fields().len());
        let mut cursor = reader.initial_position();
        for (col_pos, data_field) in self.row_type.fields().iter().enumerate() {
            let dtype = &data_field.data_type;
            if dtype.is_nullable() && reader.is_null_at(col_pos) {
                row.set_field(col_pos, Datum::Null);
                continue;

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[leekeiabstraction]

LGTM, TY for the PR!

[fresh-borzoni]

@luoyuxia Thanks for the PR
LGTM overall. One comment

we have:

impl<'a> Default for GenericRow<'a> 
 Self::new(0) 

Old code used insert, now we set:

  let mut row = GenericRow::default();                                                                                                                                                                                                               
  row.set_field(0, value);  // PANIC! index out of bounds

We should make it consistent or remove.

[luoyuxia]

Thanks all for review