chore: optimize SQL for duplicate nodes

Author: chelsea-lin

tests/unit/core/compile/sqlglot/test_compile_concat.py

@@ -24,7 +24,6 @@
 def test_compile_concat(
     scalars_types_pandas_df: pd.DataFrame, compiler_session: bigframes.Session, snapshot
 ):
-    # TODO: concat two same dataframes, which SQL does not get reused.
     # TODO: concat dataframes from a gbq table but trigger a windows compiler.
     df1 = bpd.DataFrame(scalars_types_pandas_df, session=compiler_session)
     df1 = df1[["rowindex", "int64_col", "string_col"]]