Feat: any_value hive/spark/presto closes #2053

tobymao · Aug 14, 2023 · c37abfd · c37abfd
1 parent cad6de8
commit c37abfd
Show file tree

Hide file tree

Showing 5 changed files with 18 additions and 0 deletions.
diff --git a/sqlglot/dialects/hive.py b/sqlglot/dialects/hive.py
@@ -378,6 +378,7 @@ class Generator(generator.Generator):
                 ]
             ),
             exp.Property: _property_sql,
+            exp.AnyValue: rename_func("FIRST"),
             exp.ApproxDistinct: approx_count_distinct_sql,
             exp.ArrayConcat: rename_func("CONCAT"),
             exp.ArrayJoin: lambda self, e: self.func("CONCAT_WS", e.expression, e.this),

diff --git a/sqlglot/dialects/presto.py b/sqlglot/dialects/presto.py
@@ -178,6 +178,7 @@ class Tokenizer(tokens.Tokenizer):
     class Parser(parser.Parser):
         FUNCTIONS = {
             **parser.Parser.FUNCTIONS,
+            "ARBITRARY": exp.AnyValue.from_arg_list,
             "APPROX_DISTINCT": exp.ApproxDistinct.from_arg_list,
             "APPROX_PERCENTILE": _approx_percentile,
             "BITWISE_AND": binary_from_function(exp.BitwiseAnd),
@@ -247,6 +248,7 @@ class Generator(generator.Generator):
 
         TRANSFORMS = {
             **generator.Generator.TRANSFORMS,
+            exp.AnyValue: rename_func("ARBITRARY"),
             exp.ApproxDistinct: _approx_distinct_sql,
             exp.ApproxQuantile: rename_func("APPROX_PERCENTILE"),
             exp.Array: lambda self, e: f"ARRAY[{self.expressions(e, flat=True)}]",

diff --git a/sqlglot/dialects/spark.py b/sqlglot/dialects/spark.py
@@ -56,6 +56,7 @@ class Generator(Spark2.Generator):
                 "DATEADD", e.args.get("unit") or "DAY", e.expression, e.this
             ),
         }
+        TRANSFORMS.pop(exp.AnyValue)
         TRANSFORMS.pop(exp.DateDiff)
         TRANSFORMS.pop(exp.Group)
 

diff --git a/sqlglot/expressions.py b/sqlglot/expressions.py
@@ -4047,6 +4047,7 @@ class Avg(AggFunc):
 
 
 class AnyValue(AggFunc):
+    _sql_names = ["ANY_VALUE", "FIRST"]
     arg_types = {"this": True, "having": False, "max": False}
 
 

diff --git a/tests/dialects/test_presto.py b/tests/dialects/test_presto.py
@@ -480,6 +480,19 @@ def test_unnest(self):
 
     @mock.patch("sqlglot.helper.logger")
     def test_presto(self, logger):
+        self.validate_all(
+            "ARBITRARY(x)",
+            read={
+                "presto": "ANY_VALUE(x)",
+                "hive": "FIRST(x)",
+            },
+            write={
+                "hive": "FIRST(x)",
+                "presto": "ARBITRARY(x)",
+                "spark": "ANY_VALUE(x)",
+            },
+        )
+
         self.validate_identity("SELECT * FROM x OFFSET 1 LIMIT 1")
         self.validate_identity("SELECT * FROM x OFFSET 1 FETCH FIRST 1 ROWS ONLY")
         self.validate_identity("SELECT BOOL_OR(a > 10) FROM asd AS T(a)")
Original file line number	Diff line number	Diff line change
Expand Up		@@ -4047,6 +4047,7 @@ class Avg(AggFunc):


		class AnyValue(AggFunc):
		_sql_names = ["ANY_VALUE", "FIRST"]
		arg_types = {"this": True, "having": False, "max": False}


Expand Down