feat(snowflake): FINAL/RUNNING keywords in MATCH_RECOGNIZE MEASURES (#…

…3284)

* feat(snowflake): FINAL/RUNNING keywords in MATCH_RECOGNIZE MEASURES

* Refactor parse_match_recognize_measure

Co-authored-by: Jo <46752250+georgesittas@users.noreply.github.com>

* First iteration

---------

Co-authored-by: Jo <46752250+georgesittas@users.noreply.github.com>

sqlglot/expressions.py

@@ -2225,6 +2225,13 @@ class Lateral(UDTF):
  }
 
 
+class MatchRecognizeMeasure(Expression):
+ arg_types = {
+ "this": True,
+ "window_frame": False,
+ }
+
+
 class MatchRecognize(Expression):
  arg_types = {
  "partition_by": False,

sqlglot/generator.py

@@ -2111,6 +2111,14 @@ def ordered_sql(self, expression: exp.Ordered) -> str:
 
  return f"{this}{sort_order}{nulls_sort_change}{with_fill}"
 
+ def matchrecognizemeasure_sql(self, expression: exp.MatchRecognizeMeasure) -> str:
+ window_frame = self.sql(expression, "window_frame")
+ window_frame = f"{window_frame} " if window_frame else ""
+
+ this = self.sql(expression, "this")
+
+ return f"{window_frame}{this}"
+
  def matchrecognize_sql(self, expression: exp.MatchRecognize) -> str:
  partition = self.partition_by_sql(expression)
  order = self.sql(expression, "order")

sqlglot/parser.py

@@ -2738,6 +2738,13 @@ def _parse_from(
  exp.From, comments=self._prev_comments, this=self._parse_table(joins=joins)
  )
 
+ def _parse_match_recognize_measure(self) -> exp.MatchRecognizeMeasure:
+ return self.expression(
+ exp.MatchRecognizeMeasure,
+ window_frame=self._match_texts(("FINAL", "RUNNING")) and self._prev.text.upper(),
+ this=self._parse_expression(),
+ )
+
  def _parse_match_recognize(self) -> t.Optional[exp.MatchRecognize]:
  if not self._match(TokenType.MATCH_RECOGNIZE):
  return None
@@ -2746,7 +2753,12 @@ def _parse_match_recognize(self) -> t.Optional[exp.MatchRecognize]:
 
  partition = self._parse_partition_by()
  order = self._parse_order()
- measures = self._parse_expressions() if self._match_text_seq("MEASURES") else None
+
+ measures = (
+ self._parse_csv(self._parse_match_recognize_measure)
+ if self._match_text_seq("MEASURES")
+ else None
+ )
 
  if self._match_text_seq("ONE", "ROW", "PER", "MATCH"):
  rows = exp.var("ONE ROW PER MATCH")

tests/dialects/test_snowflake.py

@@ -1575,38 +1575,42 @@ def test_regexp_replace(self, logger):
  )
 
  def test_match_recognize(self):
- for row in (
- "ONE ROW PER MATCH",
- "ALL ROWS PER MATCH",
- "ALL ROWS PER MATCH SHOW EMPTY MATCHES",
- "ALL ROWS PER MATCH OMIT EMPTY MATCHES",
- "ALL ROWS PER MATCH WITH UNMATCHED ROWS",
- ):
- for after in (
- "AFTER MATCH SKIP",
- "AFTER MATCH SKIP PAST LAST ROW",
- "AFTER MATCH SKIP TO NEXT ROW",
- "AFTER MATCH SKIP TO FIRST x",
- "AFTER MATCH SKIP TO LAST x",
+ for window_frame in ("", "FINAL ", "RUNNING "):
+ for row in (
+ "ONE ROW PER MATCH",
+ "ALL ROWS PER MATCH",
+ "ALL ROWS PER MATCH SHOW EMPTY MATCHES",
+ "ALL ROWS PER MATCH OMIT EMPTY MATCHES",
+ "ALL ROWS PER MATCH WITH UNMATCHED ROWS",
  ):
- self.validate_identity(
- f"""SELECT
+ for after in (
+ "AFTER MATCH SKIP",
+ "AFTER MATCH SKIP PAST LAST ROW",
+ "AFTER MATCH SKIP TO NEXT ROW",
+ "AFTER MATCH SKIP TO FIRST x",
+ "AFTER MATCH SKIP TO LAST x",
+ ):
+ with self.subTest(
+ f"MATCH_RECOGNIZE with window frame {window_frame}, rows {row}, after {after}: "
+ ):
+ self.validate_identity(
+ f"""SELECT
  *
 FROM x
 MATCH_RECOGNIZE (
  PARTITION BY a, b
  ORDER BY
  x DESC
  MEASURES
- y AS b
+ {window_frame}y AS b
  {row}
  {after}
  PATTERN (^ S1 S2*? ( {{- S3 -}} S4 )+ | PERMUTE(S1, S2){{1,2}} $)
  DEFINE
  x AS y
 )""",
- pretty=True,
- )
+  pretty=True,
+  )
 
  def test_show_users(self):
  self.validate_identity("SHOW USERS")