refactor: add sqlscalarop to the sqlglot compiler

Author: chelsea-lin

bigframes/core/compile/sqlglot/expressions/generic_ops.py

@@ -14,6 +14,7 @@
 
 from __future__ import annotations
 
+import sqlglot as sg
 import sqlglot.expressions as sge
 
 from bigframes import dtypes
@@ -80,6 +81,17 @@ def _(expr: TypedExpr) -> sge.Expression:
     return sge.BitwiseNot(this=sge.paren(expr.expr))
 
 
+@register_nary_op(ops.SqlScalarOp, pass_op=True)
+def _(*operands: TypedExpr, op: ops.SqlScalarOp) -> sge.Expression:
+    # TODO: can we include a string in the sqlglot expression without parsing?
+    return sg.parse_one(
+        op.sql_template.format(
+            *[operand.expr.sql(dialect="bigquery") for operand in operands]
+        ),
+        dialect="bigquery",
+    )
+
+
 @register_unary_op(ops.isnull_op)
 def _(expr: TypedExpr) -> sge.Expression:
     return sge.Is(this=expr.expr, expression=sge.Null())

tests/unit/core/compile/sqlglot/expressions/snapshots/test_generic_ops/test_sql_scalar_op/out.sql

@@ -0,0 +1,14 @@
+WITH `bfcte_0` AS (
+  SELECT
+    `bool_col` AS `bfcol_0`,
+    `bytes_col` AS `bfcol_1`
+  FROM `bigframes-dev`.`sqlglot_test`.`scalar_types`
+), `bfcte_1` AS (
+  SELECT
+    *,
+    CAST(`bfcol_0` AS INT64) + BYTE_LENGTH(`bfcol_1`) AS `bfcol_2`
+  FROM `bfcte_0`
+)
+SELECT
+  `bfcol_2` AS `bool_col`
+FROM `bfcte_1`

tests/unit/core/compile/sqlglot/expressions/test_generic_ops.py

@@ -261,6 +261,17 @@ def test_notnull(scalar_types_df: bpd.DataFrame, snapshot):
     snapshot.assert_match(sql, "out.sql")
 
 
+def test_sql_scalar_op(scalar_types_df: bpd.DataFrame, snapshot):
+    bf_df = scalar_types_df[["bool_col", "bytes_col"]]
+    sql = utils._apply_nary_op(
+        bf_df,
+        ops.SqlScalarOp(dtypes.INT_DTYPE, "CAST({0} AS INT64) + BYTE_LENGTH({1})"),
+        "bool_col",
+        "bytes_col",
+    )
+    snapshot.assert_match(sql, "out.sql")
+
+
 def test_map(scalar_types_df: bpd.DataFrame, snapshot):
     col_name = "string_col"
     bf_df = scalar_types_df[[col_name]]