in crates/core/src/merge/record_merger.rs [261:307]
fn test_merge_records_append_only() {
let schema = create_test_schema(false);
// First batch
let batch1 = RecordBatch::try_new(
schema.clone(),
vec![
Arc::new(StringArray::from(vec!["s1", "s1"])),
Arc::new(StringArray::from(vec!["k1", "k2"])),
Arc::new(Int32Array::from(vec![1, 2])),
Arc::new(Int32Array::from(vec![10, 20])),
],
)
.unwrap();
// Second batch
let batch2 = RecordBatch::try_new(
schema.clone(),
vec![
Arc::new(StringArray::from(vec!["s2", "s2"])),
Arc::new(StringArray::from(vec!["k1", "k3"])),
Arc::new(Int32Array::from(vec![3, 4])),
Arc::new(Int32Array::from(vec![30, 40])),
],
)
.unwrap();
let configs = create_configs("APPEND_ONLY", false, None);
let merger = RecordMerger::new(Arc::new(configs));
let merged = merger
.merge_record_batches(&schema, &[batch1, batch2])
.unwrap();
// Should contain all records in order without deduplication
assert_eq!(merged.num_rows(), 4);
let result = get_sorted_rows(&merged);
assert_eq!(
result,
vec![
("s1".to_string(), "k1".to_string(), 1, 10),
("s2".to_string(), "k1".to_string(), 3, 30),
("s1".to_string(), "k2".to_string(), 2, 20),
("s2".to_string(), "k3".to_string(), 4, 40),
]
);
}