Guard the compaction test against a no-op rewrite

rewrite_data_files commits no snapshot when there's nothing to compact, which
would silently shift the fixture's snapshot indices and make the double-count
test scan to the wrong (also non-append) snapshot while still passing. Assert
snapshots[3] really is the compaction (operation=replace) so a no-op fails loudly.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Author: smaheshwar-pltr

tests/integration/test_reads.py

@@ -51,6 +51,7 @@
 )
 from pyiceberg.schema import Schema
 from pyiceberg.table import Table
+from pyiceberg.table.snapshots import Operation
 from pyiceberg.types import (
     BinaryType,
     BooleanType,
@@ -1318,6 +1319,12 @@ def test_incremental_append_scan_does_not_double_count_compacted_files(catalog:
     # compacts them into a single rewritten file. A scan spanning the compaction must read each
     # appended row exactly once -- the rewritten file (added by the compaction, not by an append)
     # must not be picked up on top of the originals.
+    #
+    # Guard the fixture: rewrite_data_files is a no-op (no snapshot) when there's nothing to
+    # compact, which would silently shift the indices and make this test scan to the wrong
+    # snapshot. Assert snapshots[3] really is the compaction.
+    assert test_table.snapshots()[3].summary.operation == Operation.REPLACE  # type: ignore[union-attr]
+
     scan = test_table.incremental_append_scan(
         from_snapshot_id_exclusive=test_table.snapshots()[0].snapshot_id,
         to_snapshot_id_inclusive=test_table.snapshots()[3].snapshot_id,