E021-09: TRIM function (#114)

`TRIM` can be constructed in several forms: TRIM( [ [ { LEADING | TRAILING | BOTH } ] [ trim_character ] FROM ] trim_source ) If `LEADING`, `TRAILING` or `BOTH` is not provided, `BOTH` is used. If `trim_character` is not provided, a space (`' '`) is used. VALUES TRIM(' hello world '); -- COL1: hello world VALUES TRIM('a' FROM 'aaababccaa'); -- COL1: babcc VALUES TRIM(LEADING 'a' FROM 'aaababccaa'); -- COL1: babccaa VALUES TRIM(TRAILING 'a' FROM 'aaababccaa'); -- COL1: aaababcc
elliotchance · Jul 15, 2022 · ab3a5e7 · ab3a5e7
1 parent f84fb96
commit ab3a5e7
Show file tree

Hide file tree

Showing 10 changed files with 413 additions and 3 deletions.
diff --git a/docs/functions.rst b/docs/functions.rst
@@ -445,6 +445,36 @@ be used.
   VALUES SUBSTRING('Жabڣc' FROM 4 USING OCTETS);
   -- COL1: bڣc
 
+TRIM() CHARACTER VARYING
+^^^^^^^^^^^^^^^^^^^^^^^^
+
+``TRIM`` can be constructed in several forms:
+
+.. code-block:: text
+
+  TRIM(
+    [ [ { LEADING | TRAILING | BOTH } ] [ trim_character ] FROM ]
+    trim_source
+  )
+
+If ``LEADING``, ``TRAILING`` or ``BOTH`` is not provided, ``BOTH`` is used.
+
+If ``trim_character`` is not provided, a space (`' '`) is used.
+
+.. code-block:: sql
+
+  VALUES TRIM('  hello world ');
+  -- COL1: hello world
+
+  VALUES TRIM('a' FROM 'aaababccaa');
+  -- COL1: babcc
+
+  VALUES TRIM(LEADING 'a' FROM 'aaababccaa');
+  -- COL1: babccaa
+
+  VALUES TRIM(TRAILING 'a' FROM 'aaababccaa');
+  -- COL1: aaababcc
+
 ``UPPER(CHARACTER VARYING) CHARACTER VARYING`` 
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 

diff --git a/docs/sql-compliance.rst b/docs/sql-compliance.rst
@@ -13,7 +13,7 @@ Mandatory Features
 ------------------
 
 As of the latest version (or at least the version of this documentation)
-**vsql supports 54 of the 164 mandatory features** of the
+**vsql supports 56 of the 164 mandatory features** of the
 `SQL:2016 Standard <https://www.iso.org/standard/63556.html>`_.
 
 .. list-table:: Table 43 — Feature taxonomy and definition for mandatory features
@@ -70,7 +70,7 @@ As of the latest version (or at least the version of this documentation)
    * - ✅ E021-08
      - ``UPPER`` and ``LOWER`` functions
 
-   * - ❌ E021-09
+   * - ✅ E021-09
      - ``TRIM`` function
 
    * - ❌ E021-10

diff --git a/generate-grammar.py b/generate-grammar.py
@@ -515,6 +515,7 @@ def parse_tree(text):
 # parse_tree("SELECT x , y FROM ( SELECT x , y FROM t1 )")
 # parse_tree("VALUES TIMESTAMP '2022-06-30'")
 # parse_tree("INSERT INTO foo ( f1 ) VALUES ( TIMESTAMP '2022-06-30' )")
+# parse_tree("VALUES TRIM ( 'helloworld' )")
 
 for arg in sys.argv[1:]:
     print(arg)

diff --git a/grammar.bnf b/grammar.bnf
@@ -955,6 +955,28 @@
 <character value function> /* Expr */ ::=
     <character substring function>
   | <fold>
+  | <trim function>
+
+<trim function> /* Expr */ ::=
+  TRIM <left paren> <trim operands> <right paren>   -> trim
+
+<trim operands> /* Expr */ ::=
+    <trim source>                                              -> trim1
+  | FROM <trim source>                                         -> trim1
+  | <trim specification> FROM <trim source>                    -> trim2
+  | <trim character> FROM <trim source>                        -> trim3
+  | <trim specification> <trim character> FROM <trim source>   -> trim4
+
+<trim source> /* Expr */ ::=
+    <character value expression>
+
+<trim specification> /* string */ ::=
+    LEADING
+  | TRAILING
+  | BOTH
+
+<trim character> /* Expr */ ::=
+    <character value expression>
 
 <fold> /* Expr */ ::=
     UPPER <left paren> <character value expression> <right paren>   -> upper

diff --git a/tests/trim.sql b/tests/trim.sql
@@ -0,0 +1,44 @@
+EXPLAIN VALUES TRIM('aaababccaa');
+-- EXPLAIN: VALUES (COL1 CHARACTER VARYING) = ROW(TRIM(BOTH ' ' FROM 'aaababccaa'))
+
+EXPLAIN VALUES TRIM(FROM 'aaababccaa');
+-- EXPLAIN: VALUES (COL1 CHARACTER VARYING) = ROW(TRIM(BOTH ' ' FROM 'aaababccaa'))
+
+EXPLAIN VALUES TRIM(LEADING FROM 'aaababccaa');
+-- EXPLAIN: VALUES (COL1 CHARACTER VARYING) = ROW(TRIM(LEADING ' ' FROM 'aaababccaa'))
+
+EXPLAIN VALUES TRIM(TRAILING FROM 'aaababccaa');
+-- EXPLAIN: VALUES (COL1 CHARACTER VARYING) = ROW(TRIM(TRAILING ' ' FROM 'aaababccaa'))
+
+EXPLAIN VALUES TRIM(BOTH FROM 'aaababccaa');
+-- EXPLAIN: VALUES (COL1 CHARACTER VARYING) = ROW(TRIM(BOTH ' ' FROM 'aaababccaa'))
+
+EXPLAIN VALUES TRIM('a' FROM 'aaababccaa');
+-- EXPLAIN: VALUES (COL1 CHARACTER VARYING) = ROW(TRIM(BOTH 'a' FROM 'aaababccaa'))
+
+EXPLAIN VALUES TRIM(LEADING 'a' FROM 'aaababccaa');
+-- EXPLAIN: VALUES (COL1 CHARACTER VARYING) = ROW(TRIM(LEADING 'a' FROM 'aaababccaa'))
+
+VALUES TRIM('aaababccaa');
+-- COL1: aaababccaa
+
+VALUES TRIM(FROM 'aaababccaa');
+-- COL1: aaababccaa
+
+VALUES TRIM(LEADING FROM 'aaababccaa');
+-- COL1: aaababccaa
+
+VALUES TRIM(TRAILING FROM 'aaababccaa');
+-- COL1: aaababccaa
+
+VALUES TRIM(BOTH FROM 'aaababccaa');
+-- COL1: aaababccaa
+
+VALUES TRIM('a' FROM 'aaababccaa');
+-- COL1: babcc
+
+VALUES TRIM(LEADING 'a' FROM 'aaababccaa');
+-- COL1: babccaa
+
+VALUES TRIM(TRAILING 'a' FROM 'aaababccaa');
+-- COL1: aaababcc
diff --git a/vsql/ast.v b/vsql/ast.v
@@ -37,6 +37,7 @@ type Expr = BetweenExpr
 	| RowExpr
 	| SimilarExpr
 	| SubstringExpr
+	| TrimExpr
 	| UnaryExpr
 	| Value
 
@@ -103,6 +104,9 @@ fn (e Expr) pstr(params map[string]Value) string {
 		SubstringExpr {
 			e.pstr(params)
 		}
+		TrimExpr {
+			e.pstr(params)
+		}
 		UnaryExpr {
 			e.pstr(params)
 		}
@@ -553,3 +557,17 @@ fn (e SubstringExpr) pstr(params map[string]Value) string {
 
 	return s + ' USING $e.using)'
 }
+
+struct TrimExpr {
+	specification string // LEADING, TRAILING or BOTH
+	character     Expr   // NoExpr when missing
+	source        Expr
+}
+
+fn (e TrimExpr) str() string {
+	return e.pstr(map[string]Value{})
+}
+
+fn (e TrimExpr) pstr(params map[string]Value) string {
+	return 'TRIM($e.specification ${e.character.pstr(params)} FROM ${e.source.pstr(params)})'
+}
diff --git a/vsql/eval.v b/vsql/eval.v
@@ -144,7 +144,7 @@ fn eval_as_type(conn &Connection, data Row, e Expr, params map[string]Value) ?Ty
 		LocalTimestampExpr {
 			return new_type('TIMESTAMP WITHOUT TIME ZONE', 0)
 		}
-		SubstringExpr {
+		SubstringExpr, TrimExpr {
 			return new_type('CHARACTER VARYING', 0)
 		}
 	}
@@ -233,6 +233,9 @@ fn eval_as_value(conn &Connection, data Row, e Expr, params map[string]Value) ?V
 
 			return new_timestamp_value(now.strftime('%Y-%m-%d ') + time_value(conn, e.prec, false))
 		}
+		TrimExpr {
+			return eval_trim(conn, data, e, params)
+		}
 	}
 }
 
@@ -324,6 +327,23 @@ fn eval_null(conn &Connection, data Row, e NullExpr, params map[string]Value) ?V
 	return new_boolean_value(value.is_null())
 }
 
+fn eval_trim(conn &Connection, data Row, e TrimExpr, params map[string]Value) ?Value {
+	source := eval_as_value(conn, data, e.source, params)?
+	character := eval_as_value(conn, data, e.character, params)?
+
+	if e.specification == 'LEADING' {
+		return new_varchar_value(source.string_value.trim_left(character.string_value),
+			0)
+	}
+
+	if e.specification == 'TRAILING' {
+		return new_varchar_value(source.string_value.trim_right(character.string_value),
+			0)
+	}
+
+	return new_varchar_value(source.string_value.trim(character.string_value), 0)
+}
+
 fn eval_like(conn &Connection, data Row, e LikeExpr, params map[string]Value) ?Value {
 	left := eval_as_value(conn, data, e.left, params)?
 	right := eval_as_value(conn, data, e.right, params)?

diff --git a/vsql/expr.v b/vsql/expr.v
@@ -63,6 +63,11 @@ fn expr_is_agg(conn &Connection, e Expr) ?bool {
 				return nested_agg_unsupported(e)
 			}
 		}
+		TrimExpr {
+			if expr_is_agg(conn, e.source)? || expr_is_agg(conn, e.character)? {
+				return nested_agg_unsupported(e)
+			}
+		}
 	}
 
 	return false
@@ -136,5 +141,9 @@ fn resolve_identifiers(e Expr, tables map[string]Table) ?Expr {
 			// These don't have any Expr properties to recurse.
 			return e
 		}
+		TrimExpr {
+			return TrimExpr{e.specification, resolve_identifiers(e.character, tables)?, resolve_identifiers(e.source,
+				tables)?}
+		}
 	}
 }