#3649 fix(targets): Reset duplicate-merged tally between batches Pull request opened by edgarrmondragon Summary •

Sink.mark_drained()

reset

_batch_records_read

to 0 but never reset

_batch_dupe_records_merged

• On each subsequent drain,

tally_record_written()

subtracted all accumulated duplicates (not just those from the current batch), making

_total_records_written

increasingly wrong • Fixed by resetting

_batch_dupe_records_merged = 0

alongside

_batch_records_read = 0

in

mark_drained()

• Added parametrized regression test that verifies all records are written across multiple drain cycles when

batch_size_rows < total_records

Test plan •

uv run pytest tests/core/test_target_base.py -v

Summary by Sourcery Reset duplicate-merge batch counters correctly between drain cycles and add regression coverage to ensure all records are written when using row-based batching. Bug Fixes: • Reset the per-batch duplicate-merged counter when draining so total records written is tallied correctly across multiple batches. Tests: • Add a parametrized regression test to verify no records are lost and batch counts are correct when batch_size_rows is smaller than the total number of records. meltano/sdk