feat: Support for RPAD & LPAD functions (#3757)

* feat: Support for RPAD & LPAD functions

* Switch to boolean-based kind

* Refactor boolean arg

Author: VaggelisD

sqlglot/dialects/duckdb.py

@@ -389,6 +389,7 @@ class Generator(generator.Generator):
         SUPPORTS_TO_NUMBER = False
         COPY_HAS_INTO_KEYWORD = False
         STAR_EXCEPT = "EXCLUDE"
+        PAD_FILL_PATTERN_IS_REQUIRED = True
 
         TRANSFORMS = {
             **generator.Generator.TRANSFORMS,

sqlglot/dialects/hive.py

@@ -447,6 +447,7 @@ class Generator(generator.Generator):
         SUPPORTS_TO_NUMBER = False
         WITH_PROPERTIES_PREFIX = "TBLPROPERTIES"
         PARSE_JSON_NAME = None
+        PAD_FILL_PATTERN_IS_REQUIRED = True
 
         EXPRESSIONS_WITHOUT_NESTED_CTES = {
             exp.Insert,

sqlglot/dialects/mysql.py

@@ -690,6 +690,7 @@ class Generator(generator.Generator):
         JSON_KEY_VALUE_PAIR_SEP = ","
         SUPPORTS_TO_NUMBER = False
         PARSE_JSON_NAME = None
+        PAD_FILL_PATTERN_IS_REQUIRED = True
 
         TRANSFORMS = {
             **generator.Generator.TRANSFORMS,

sqlglot/dialects/presto.py

@@ -352,6 +352,7 @@ class Generator(generator.Generator):
         SUPPORTS_TO_NUMBER = False
         HEX_FUNC = "TO_HEX"
         PARSE_JSON_NAME = "JSON_PARSE"
+        PAD_FILL_PATTERN_IS_REQUIRED = True
 
         PROPERTIES_LOCATION = {
             **generator.Generator.PROPERTIES_LOCATION,

sqlglot/dialects/spark.py

@@ -131,6 +131,7 @@ def _parse_generated_as_identity(
 
     class Generator(Spark2.Generator):
         SUPPORTS_TO_NUMBER = True
+        PAD_FILL_PATTERN_IS_REQUIRED = False
 
         TYPE_MAPPING = {
             **Spark2.Generator.TYPE_MAPPING,

sqlglot/expressions.py

@@ -4798,6 +4798,11 @@ class List(Func):
     is_var_len_args = True
 
 
+# String pad, kind True -> LPAD, False -> RPAD
+class Pad(Func):
+    arg_types = {"this": True, "expression": True, "fill_pattern": False, "is_left": True}
+
+
 # https://docs.snowflake.com/en/sql-reference/functions/to_char
 # https://docs.oracle.com/en/database/oracle/oracle-database/23/sqlrf/TO_CHAR-number.html
 class ToChar(Func):

sqlglot/generator.py

@@ -375,6 +375,9 @@ class Generator(metaclass=_Generator):
     # Whether to quote the generated expression of exp.JsonPath
     QUOTE_JSON_PATH = True
 
+    # Whether the text pattern/fill (3rd) parameter of RPAD()/LPAD() is optional (defaults to space)
+    PAD_FILL_PATTERN_IS_REQUIRED = False
+
     # The name to generate for the JSONPath expression. If `None`, only `this` will be generated
     PARSE_JSON_NAME: t.Optional[str] = "PARSE_JSON"
 
@@ -4048,3 +4051,12 @@ def changes_sql(self, expression: exp.Changes) -> str:
         end = f"{self.seg('')}{end}" if end else ""
 
         return f"CHANGES ({information}){at_before}{end}"
+
+    def pad_sql(self, expression: exp.Pad) -> str:
+        prefix = "L" if expression.args.get("is_left") else "R"
+
+        fill_pattern = self.sql(expression, "fill_pattern") or None
+        if not fill_pattern and self.PAD_FILL_PATTERN_IS_REQUIRED:
+            fill_pattern = "' '"
+
+        return self.func(f"{prefix}PAD", expression.this, expression.expression, fill_pattern)

sqlglot/parser.py

@@ -108,6 +108,15 @@ def build_mod(args: t.List) -> exp.Mod:
     return exp.Mod(this=this, expression=expression)
 
 
+def build_pad(args: t.List, is_left: bool = True):
+    return exp.Pad(
+        this=seq_get(args, 0),
+        expression=seq_get(args, 1),
+        fill_pattern=seq_get(args, 2),
+        is_left=is_left,
+    )
+
+
 class _Parser(type):
     def __new__(cls, clsname, bases, attrs):
         klass = super().__new__(cls, clsname, bases, attrs)
@@ -159,7 +168,11 @@ class Parser(metaclass=_Parser):
         "LOG2": lambda args: exp.Log(this=exp.Literal.number(2), expression=seq_get(args, 0)),
         "LOG10": lambda args: exp.Log(this=exp.Literal.number(10), expression=seq_get(args, 0)),
         "LOWER": build_lower,
+        "LPAD": lambda args: build_pad(args),
+        "LEFTPAD": lambda args: build_pad(args),
         "MOD": build_mod,
+        "RPAD": lambda args: build_pad(args, is_left=False),
+        "RIGHTPAD": lambda args: build_pad(args, is_left=False),
         "SCOPE_RESOLUTION": lambda args: exp.ScopeResolution(expression=seq_get(args, 0))
         if len(args) != 2
         else exp.ScopeResolution(this=seq_get(args, 0), expression=seq_get(args, 1)),

tests/dialects/test_dialect.py

@@ -2573,3 +2573,33 @@ def test_reserved_keywords(self):
             """SELECT partition.d FROM t PARTITION (d)""",
             """SELECT partition.d FROM t AS PARTITION(d)""",
         )
+
+    def test_string_functions(self):
+        for pad_func in ("LPAD", "RPAD"):
+            ch_alias = "LEFTPAD" if pad_func == "LPAD" else "RIGHTPAD"
+            for fill_pattern in ("", ", ' '"):
+                with self.subTest(f"Testing {pad_func}() with pattern {fill_pattern}"):
+                    self.validate_all(
+                        f"SELECT {pad_func}('bar', 5{fill_pattern})",
+                        read={
+                            "snowflake": f"SELECT {pad_func}('bar', 5{fill_pattern})",
+                            "databricks": f"SELECT {pad_func}('bar', 5{fill_pattern})",
+                            "spark": f"SELECT {pad_func}('bar', 5{fill_pattern})",
+                            "postgres": f"SELECT {pad_func}('bar', 5{fill_pattern})",
+                            "clickhouse": f"SELECT {ch_alias}('bar', 5{fill_pattern})",
+                        },
+                        write={
+                            "": f"SELECT {pad_func}('bar', 5{fill_pattern})",
+                            "spark": f"SELECT {pad_func}('bar', 5{fill_pattern})",
+                            "postgres": f"SELECT {pad_func}('bar', 5{fill_pattern})",
+                            "clickhouse": f"SELECT {pad_func}('bar', 5{fill_pattern})",
+                            "snowflake": f"SELECT {pad_func}('bar', 5{fill_pattern})",
+                            "databricks": f"SELECT {pad_func}('bar', 5{fill_pattern})",
+                            "duckdb": f"SELECT {pad_func}('bar', 5, ' ')",
+                            "mysql": f"SELECT {pad_func}('bar', 5, ' ')",
+                            "hive": f"SELECT {pad_func}('bar', 5, ' ')",
+                            "spark2": f"SELECT {pad_func}('bar', 5, ' ')",
+                            "presto": f"SELECT {pad_func}('bar', 5, ' ')",
+                            "trino": f"SELECT {pad_func}('bar', 5, ' ')",
+                        },
+                    )