feat(duckdb): Add transpilation support for SUBSTR-SUBSTRING functions

Author: fivetran-amrutabhimsenayachit

sqlglot-integration-tests

@@ -220,7 +220,7 @@ class Stuff(Expression, Func):
 
 class Substring(Expression, Func):
     _sql_names = ["SUBSTRING", "SUBSTR"]
-    arg_types = {"this": True, "start": False, "length": False}
+    arg_types = {"this": True, "start": False, "length": False, "zero_start": False}
 
 
 class SubstringIndex(Expression, Func):

sqlglot/expressions/string.py

@@ -2504,6 +2504,35 @@ def strposition_sql(self, expression: exp.StrPosition) -> str:
 
         return strposition_sql(self, expression)
 
+    def substring_sql(self, expression: exp.Substring) -> str:
+        if expression.args.get("zero_start"):
+            start = expression.args.get("start")
+            length = expression.args.get("length")
+
+            new_start = (
+                exp.Literal.number(1)
+                if start is not None and start.is_number and start.to_py() == 0
+                else exp.If(this=start.eq(0), true=exp.Literal.number(1), false=start.copy())
+                if start is not None and not start.is_number
+                else None
+            )
+            new_length = (
+                exp.Literal.number(0)
+                if length is not None and length.is_number and length.to_py() < 0
+                else exp.If(this=length.copy() < 0, true=exp.Literal.number(0), false=length.copy())
+                if length is not None and not length.is_number
+                else None
+            )
+
+            if new_start is not None or new_length is not None:
+                expression = expression.copy()
+                if new_start is not None:
+                    expression.set("start", new_start)
+                if new_length is not None:
+                    expression.set("length", new_length)
+
+        return self.function_fallback_sql(expression)
+
     def strtotime_sql(self, expression: exp.StrToTime) -> str:
         # Check if target_type requires TIMESTAMPTZ (for LTZ/TZ variants)
         target_type = expression.args.get("target_type")

sqlglot/generators/duckdb.py

@@ -726,6 +726,7 @@ class SnowflakeParser(parser.Parser):
         "OBJECT_CONSTRUCT_KEEP_NULL": lambda self: self._parse_json_object(),
         "LISTAGG": lambda self: self._parse_string_agg(),
         "SEMANTIC_VIEW": lambda self: self._parse_semantic_view(),
+        "SUBSTR": lambda self: self._parse_substring(),
     }
     FUNCTION_PARSERS = {k: v for k, v in FUNCTION_PARSERS.items() if k != "TRIM"}
 
@@ -1284,6 +1285,11 @@ def _parse_position(self, haystack_first: bool = False) -> exp.StrPosition:
         result.set("clamp_position", True)
         return result
 
+    def _parse_substring(self) -> exp.Substring:
+        result = super()._parse_substring()
+        result.set("zero_start", True)
+        return result
+
     def _parse_window(self, this: exp.Expr | None, alias: bool = False) -> exp.Expr | None:
         if isinstance(this, exp.NthValue):
             if self._match_text_seq("FROM"):