implemented insert statement

frontend-hyrise/sample-queries.sql

@@ -11,7 +11,5 @@ SELECT name, city, grade FROM (SELECT * FROM students WHERE city = 'Potsdam') t1
 SELECT city, AVG(grade) AS average|, MIN(grade) AS best, MAX(grade) AS worst FROM students GROUP BY city;
 # UNION
 SELECT * FROM students WHERE grade = 1.3 UNION SELECT * FROM students WHERE grade = 3.7;
-# SIMPLE JOIN
+# JOIN
-SELECT * FROM companies JOIN employees ON company_id = employee_company_id;
+SELECT * FROM companies JOIN employees ON company_id = employee_company_id;
-# HASH JOIN
-SELECT * FROM companies HASH JOIN employees ON company_id = employee_company_id;

run_tests.sh

@@ -12,7 +12,9 @@ echo "\n\n"
 
 # ./bin/analysis "SELECT t1.a AS id, t1.b, t2.c FROM \"tbl\" AS t1 JOIN (SELECT * FROM foo JOIN bar ON foo.id = bar.id) t2 ON t1.a = t2.b WHERE (t1.b OR NOT t1.a) AND t2.c = 12.5"
 # ./bin/analysis "CREATE TABLE \"table\" FROM TBL FILE 'students.tbl'"
-./bin/analysis "SELECT * FROM t1 UNION (SELECT abc AS t FROM t2 ORDER BY col3 LIMIT 10) ORDER BY col1;"
+./bin/analysis "SELECT * FROM t1 UNION SELECT abc AS t FROM t2 ORDER BY col3 LIMIT 10;"
+./bin/analysis "INSERT INTO students (name, city, age) VALUES ('Max', 'Musterhausen', 5);"
+./bin/analysis "INSERT INTO students (name, city) SELECT * FROM employees;"
 
 echo "\n\n"
 

src/lib/InsertStatement.h

@@ -0,0 +1,36 @@
+#ifndef __INSERT_STATEMENT_H__
+#define __INSERT_STATEMENT_H__
+
+#include "Statement.h"
+#include "SelectStatement.h"
+
+namespace hsql {
+
+
+struct InsertStatement : Statement {
+	typedef enum {
+		kInsertValues,
+		kInsertSelect
+	} Type;
+
+	InsertStatement() :
+		Statement(kStmtInsert),
+		table_name(NULL),
+		columns(NULL),
+		values(NULL),
+		select(NULL) {}
+	
+	virtual ~InsertStatement(); // defined in destructors.cpp
+
+	Type insert_type;
+	const char* table_name;
+	List<char*>* columns;
+	List<Expr*>* values;
+	SelectStatement* select;
+};
+
+
+
+
+} // namsepace hsql
+#endif

src/lib/Statement.h

@@ -14,9 +14,9 @@ typedef enum {
 	kStmtError, // Unused
 	kStmtSelect,
 	kStmtImport,
+	kStmtInsert,
 	// Following types are planned but not supported yet
 	kStmtDelete,
-	kStmtInsert,
 	kStmtCreate,
 	kStmtDrop,
 	kStmtExport,

src/lib/destructors.cpp

@@ -31,6 +31,13 @@ CreateStatement::~CreateStatement() {
 	delete table_name;
 }
 
+InsertStatement::~InsertStatement() {
+	delete table_name;
+	delete select;
+}
+
+
+
 OrderDescription::~OrderDescription() {
 	delete expr;
 }

src/lib/sqlhelper.cpp

@@ -120,5 +120,28 @@ void printCreateStatementInfo(CreateStatement* stmt, uint num_indent) {
   inprint(stmt->file_path, num_indent+1);
 }
 
+void printInsertStatementInfo(InsertStatement* stmt, uint num_indent) {
+  inprint("InsertStatment", num_indent);
+  inprint(stmt->table_name, num_indent+1);
+  if (stmt->columns != NULL) {
+    inprint("Columns", num_indent+1);
+    for (char* col_name : stmt->columns->vector()) {
+      inprint(col_name, num_indent+2);
+    }
+  }
+  switch (stmt->insert_type) {
+    case InsertStatement::kInsertValues:
+      inprint("Values", num_indent+1);
+      for (Expr* expr : stmt->values->vector()) {
+        printExpression(expr, num_indent+2);
+      }
+      break;
+    case InsertStatement::kInsertSelect:
+      printSelectStatementInfo(stmt->select, num_indent+1);
+      break;
+  }
+
+}
+
 
 } // namespace hsql

src/lib/sqlhelper.h

@@ -9,6 +9,7 @@ namespace hsql {
 
 void printSelectStatementInfo(SelectStatement* stmt, uint num_indent);
 void printImportStatementInfo(ImportStatement* stmt, uint num_indent);
+void printInsertStatementInfo(InsertStatement* stmt, uint num_indent);
 void printCreateStatementInfo(CreateStatement* stmt, uint num_indent);
 
 } // namespace hsql

src/lib/sqllib.h

@@ -5,5 +5,6 @@
 #include "SelectStatement.h"
 #include "ImportStatement.h"
 #include "CreateStatement.h"
+#include "InsertStatement.h"
 
 #endif

src/parser/bison_parser.y

@@ -82,6 +82,7 @@ typedef void* yyscan_t;
 	hsql::SelectStatement* select_stmt;
 	hsql::ImportStatement* import_stmt;
 	hsql::CreateStatement* create_stmt;
+	hsql::InsertStatement* insert_stmt;
 
 	hsql::TableRef* table;
 	hsql::Expr* expr;
@@ -106,13 +107,17 @@ typedef void* yyscan_t;
 %token <uval> NOTEQUALS LESSEQ GREATEREQ
 
 /* SQL Keywords */
-%token DATABASE DISTINCT BETWEEN CONTROL NATURAL COLUMN
+%token INTERSECT TEMPORARY DISTINCT RESTRICT TRUNCATE
-%token CREATE DELETE EXISTS HAVING IMPORT INSERT ISNULL
+%token ANALYZE BETWEEN CASCADE COLUMNS CONTROL DEFAULT
-%token OFFSET RENAME SELECT UNLOAD UPDATE ALTER CROSS GROUP
+%token EXPLAIN HISTORY NATURAL PRIMARY SCHEMAS SPATIAL
-%token INDEX INNER LIMIT ORDER OUTER RADIX RIGHT TABLE UNION
+%token BEFORE COLUMN CREATE DELETE DIRECT ESCAPE EXCEPT
-%token USING WHERE DESC DROP FILE FROM HASH INTO JOIN LEFT
+%token EXISTS GLOBAL HAVING IMPORT INSERT ISNULL OFFSET
-%token LIKE LOAD NULL SCAN ALL AND ASC CSV NOT TBL TOP AS BY
+%token RENAME SCHEMA SELECT SORTED TABLES UNIQUE UNLOAD
-%token IF IN IS ON OR
+%token UPDATE VALUES AFTER ALTER CROSS GROUP INDEX INNER
+%token LIMIT LOCAL MINUS ORDER OUTER RIGHT TABLE UNION USING
+%token WHERE CALL DESC DROP FILE FROM FULL HASH INTO JOIN
+%token LEFT LIKE LOAD NULL PLAN SHOW WITH ADD ALL AND ASC
+%token CSV FOR KEY NOT SET TBL TOP AS BY IF IN IS ON OR TO
 
 
 /*********************************
@@ -120,9 +125,10 @@ typedef void* yyscan_t;
  *********************************/
 %type <stmt_list>	statement_list
 %type <statement> 	statement
-%type <select_stmt> select_statement select_ref select_with_paren select_no_paren select_clause
+%type <select_stmt> select_statement select_with_paren select_no_paren select_clause
 %type <import_stmt> import_statement
 %type <create_stmt> create_statement
+%type <insert_stmt> insert_statement
 %type <sval> 		table_name opt_alias alias file_path
 %type <bval> 		opt_not_exists
 %type <table> 		from_clause table_ref table_ref_atomic table_ref_name
@@ -130,12 +136,13 @@ typedef void* yyscan_t;
 %type <expr> 		expr scalar_expr unary_expr binary_expr function_expr star_expr expr_alias
 %type <expr> 		column_name literal int_literal num_literal string_literal
 %type <expr> 		comp_expr opt_where join_condition
-%type <expr_list> 	expr_list opt_group select_list
+%type <expr_list> 	expr_list opt_group select_list literal_list
 %type <table_list> 	table_ref_commalist
 %type <order>		opt_order
 %type <limit>		opt_limit
 %type <order_type>	opt_order_type
 %type <uval>		import_file_type opt_join_type
+%type <slist>		ident_commalist opt_column_list
 
 /******************************
  ** Token Precedence and Associativity
@@ -165,9 +172,7 @@ typedef void* yyscan_t;
  ** Section 3: Grammar Definition
  *********************************/
 
-
 // Defines our general input.
-// TODO: Support list of statements
 input:
 		statement_list opt_semicolon { *result = $1; }
 	;
@@ -178,13 +183,11 @@ statement_list:
 	|	statement_list ';' statement { $1->push_back($3); $$ = $1; }
 	;
 
-
-// All types of statements
-// TODO: insert, delete, etc...
 statement:
 		select_statement { $$ = $1; }
 	|	import_statement { $$ = $1; }
 	|	create_statement { $$ = $1; }
+	|	insert_statement { $$ = $1; }
 	;
 
 
@@ -228,6 +231,33 @@ opt_not_exists:
 	|	/* empty */ { $$ = false; }
 	;
 
+
+/******************************
+ ** Insert Statement
+ ******************************/
+insert_statement:
+		INSERT INTO table_name opt_column_list VALUES '(' literal_list ')' {
+			$$ = new InsertStatement();
+			$$->insert_type = InsertStatement::kInsertValues;
+			$$->table_name = $3;
+			$$->columns = $4;
+			$$->values = $7;
+		}
+	|	INSERT INTO table_name opt_column_list select_no_paren {
+			$$ = new InsertStatement();
+			$$->insert_type = InsertStatement::kInsertSelect;
+			$$->table_name = $3;
+			$$->columns = $4;
+			$$->select = $5;
+		}
+	;
+
+
+opt_column_list:
+		'(' ident_commalist ')' { $$ = $2; }
+	|	/* empty */ { $$ = NULL; }
+	;
+
 /******************************
  ** Select Statement
  ******************************/
@@ -248,18 +278,21 @@ select_no_paren:
 			$$->order = $2;
 			$$->limit = $3;
 		}
-	|	select_ref UNION select_ref opt_order opt_limit {
+	|	select_clause set_operator select_clause opt_order opt_limit {
+			// TODO: allow multiple unions (through linked list)
+			// TODO: capture type of set_operator
+			// TODO: might overwrite order and limit of first select here
 			$$ = $1;
 			$$->union_select = $3;
-			// TODO: might overwrite order and limit of first select here
 			$$->order = $4;
 			$$->limit = $5;
 		}
 	;
 
-select_ref:
+set_operator:
-		select_clause
+		UNION
-	|	select_with_paren
+	|	INTERSECT
+	|	EXCEPT
 	;
 
 select_clause:
@@ -321,6 +354,11 @@ expr_list:
 	|	expr_list ',' expr_alias { $1->push_back($3); $$ = $1; }
 	;
 
+literal_list:
+		literal { $$ = new List<Expr*>($1); }
+	|	literal_list ',' literal { $1->push_back($3); $$ = $1; }
+	;
+
 expr_alias:
 		expr opt_alias {
 			$$ = $1;
@@ -512,6 +550,11 @@ opt_semicolon:
 	;
 
 
+ident_commalist:
+		IDENTIFIER { $$ = new List<char*>($1); }
+	|	ident_commalist ',' IDENTIFIER { $1->push_back($3); $$ = $1; }
+	;
+
 %%
 /*********************************
  ** Section 4: Additional C code

src/parser/flex_lexer.l

@@ -57,42 +57,68 @@
 [ \t\n]+	/* skip whitespace */;
 
 
-DATABASE	TOKEN(DATABASE)
+INTERSECT	TOKEN(INTERSECT)
+TEMPORARY	TOKEN(TEMPORARY)
 DISTINCT	TOKEN(DISTINCT)
+RESTRICT	TOKEN(RESTRICT)
+TRUNCATE	TOKEN(TRUNCATE)
+ANALYZE		TOKEN(ANALYZE)
 BETWEEN		TOKEN(BETWEEN)
+CASCADE		TOKEN(CASCADE)
+COLUMNS		TOKEN(COLUMNS)
 CONTROL		TOKEN(CONTROL)
+DEFAULT		TOKEN(DEFAULT)
+EXPLAIN		TOKEN(EXPLAIN)
+HISTORY		TOKEN(HISTORY)
 NATURAL		TOKEN(NATURAL)
+PRIMARY		TOKEN(PRIMARY)
+SCHEMAS		TOKEN(SCHEMAS)
+SPATIAL		TOKEN(SPATIAL)
+BEFORE		TOKEN(BEFORE)
 COLUMN		TOKEN(COLUMN)
 CREATE		TOKEN(CREATE)
 DELETE		TOKEN(DELETE)
+DIRECT		TOKEN(DIRECT)
+ESCAPE		TOKEN(ESCAPE)
+EXCEPT		TOKEN(EXCEPT)
 EXISTS		TOKEN(EXISTS)
+GLOBAL		TOKEN(GLOBAL)
 HAVING		TOKEN(HAVING)
 IMPORT		TOKEN(IMPORT)
 INSERT		TOKEN(INSERT)
 ISNULL		TOKEN(ISNULL)
 OFFSET		TOKEN(OFFSET)
 RENAME		TOKEN(RENAME)
+SCHEMA		TOKEN(SCHEMA)
 SELECT		TOKEN(SELECT)
+SORTED		TOKEN(SORTED)
+TABLES		TOKEN(TABLES)
+UNIQUE		TOKEN(UNIQUE)
 UNLOAD		TOKEN(UNLOAD)
 UPDATE		TOKEN(UPDATE)
+VALUES		TOKEN(VALUES)
+AFTER		TOKEN(AFTER)
 ALTER		TOKEN(ALTER)
 CROSS		TOKEN(CROSS)
 GROUP		TOKEN(GROUP)
 INDEX		TOKEN(INDEX)
 INNER		TOKEN(INNER)
 LIMIT		TOKEN(LIMIT)
+LOCAL		TOKEN(LOCAL)
+MINUS		TOKEN(MINUS)
 ORDER		TOKEN(ORDER)
 OUTER		TOKEN(OUTER)
-RADIX		TOKEN(RADIX)
 RIGHT		TOKEN(RIGHT)
 TABLE		TOKEN(TABLE)
 UNION		TOKEN(UNION)
 USING		TOKEN(USING)
 WHERE		TOKEN(WHERE)
+CALL		TOKEN(CALL)
 DESC		TOKEN(DESC)
 DROP		TOKEN(DROP)
 FILE		TOKEN(FILE)
 FROM		TOKEN(FROM)
+FULL		TOKEN(FULL)
 HASH		TOKEN(HASH)
 INTO		TOKEN(INTO)
 JOIN		TOKEN(JOIN)
@@ -100,12 +126,18 @@ LEFT		TOKEN(LEFT)
 LIKE		TOKEN(LIKE)
 LOAD		TOKEN(LOAD)
 NULL		TOKEN(NULL)
-SCAN		TOKEN(SCAN)
+PLAN		TOKEN(PLAN)
+SHOW		TOKEN(SHOW)
+WITH		TOKEN(WITH)
+ADD			TOKEN(ADD)
 ALL			TOKEN(ALL)
 AND			TOKEN(AND)
 ASC			TOKEN(ASC)
 CSV			TOKEN(CSV)
+FOR			TOKEN(FOR)
+KEY			TOKEN(KEY)
 NOT			TOKEN(NOT)
+SET			TOKEN(SET)
 TBL			TOKEN(TBL)
 TOP			TOKEN(TOP)
 AS			TOKEN(AS)
@@ -115,6 +147,7 @@ IN			TOKEN(IN)
 IS			TOKEN(IS)
 ON			TOKEN(ON)
 OR			TOKEN(OR)
+TO			TOKEN(TO)
 
 
 "<>" 		TOKEN(NOTEQUALS)

src/parser/keywordlist_generator.py

@@ -5,7 +5,7 @@ import math
 with open("sql_keywords.txt", 'r') as fh:
 	keywords = [line.strip() for line in fh.readlines() if not line.strip().startswith("//") and len(line.strip()) > 0]
 
-	keywords = sorted(keywords) # Sort by name
+	keywords = sorted(set(keywords)) # Sort by name
 	keywords = sorted(keywords, key=lambda x: len(x), reverse=True) # Sort by length
 
 	#################

src/parser/sql_keywords.txt

@@ -1,6 +1,7 @@
 // Possible source for more tokens https://www.sqlite.org/lang_keywords.html
 
-// Select statement
+//////////////////////////
+// Select Statement
 SELECT
 TOP
 FROM
@@ -16,6 +17,9 @@ DISTINCT
 OFFSET
 UNION
 ALL
+EXCEPT
+MINUS
+INTERSECT
 
 // Join clause
 JOIN
@@ -24,42 +28,92 @@ INNER
 OUTER
 LEFT
 RIGHT
+FULL
 CROSS
 USING
 NATURAL
-HASH
+// Select Statement
-RADIX
+//////////////////////
-SCAN
+// Data Definition
-
-// Create statement
 CREATE
 TABLE
-DATABASE
+SCHEMA
 INDEX
 IF
+NOT
 EXISTS
+GLOBAL
+LOCAL
+TEMPORARY
+WITH
+UNIQUE
+
+INDEX
+UNIQUE
+HASH
+SPATIAL
+PRIMARY
+KEY
+ON
+
+DROP
+TABLE
+SCHEMA
+RESTRICT
+CASCADE
+
+ALTER
+ADD
+COLUMN
+BEFORE
+AFTER
+// Data Definition
+////////////////////////
+// Data Manipulation
+INSERT
+VALUES
+SET
+DIRECT
+SORTED
 
-// Import statement
 IMPORT
 CSV
 FILE
 TBL
 CONTROL
 
-// other statements
-ALTER
-RENAME
-DROP
-DELETE
-INSERT
-LOAD
 UPDATE
+
+DELETE
+HISTORY
+
+TRUNCATE
+
+
+// other statements
+RENAME
+TO
+DELETE
+EXPLAIN
+PLAN
+ANALYZE
+LOAD
 UNLOAD
 
+SHOW
+SCHEMAS
+TABLES
+COLUMNS
+
 // misc.
 COLUMN
 INTO
 AS
+SET
+DEFAULT
+CALL
+FOR
+
 
 // Expressions
 NOT
@@ -70,4 +124,5 @@ LIKE
 IN
 IS
 ISNULL
-BETWEEN
+BETWEEN
+ESCAPE

src/sql_analysis.cpp

@@ -30,6 +30,7 @@ int main(int argc, char *argv[]) {
                 case kStmtSelect: printSelectStatementInfo((SelectStatement*) stmt, 1); break;
                 case kStmtImport: printImportStatementInfo((ImportStatement*) stmt, 1); break;
                 case kStmtCreate: printCreateStatementInfo((CreateStatement*) stmt, 1); break;
+                case kStmtInsert: printInsertStatementInfo((InsertStatement*) stmt, 1); break;
                 default:
                     fprintf(stderr, "\tStatement Type %u. No detailed print method available.\n", stmt->type);
                     break;

test/valid_queries.sql

@@ -5,9 +5,8 @@ SELECT col1 AS myname, col2, 'test' FROM "table", foo AS t WHERE age > 12 AND zi
 SELECT * from "table" JOIN table2 ON a = b WHERE (b OR NOT a) AND a = 12.5
 (SELECT a FROM foo WHERE a > 12 OR b > 3 AND c NOT LIKE 's%' LIMIT 10);
 SELECT * FROM "table" LIMIT 10 OFFSET 10;
-SELECT * FROM t1 UNION SELECT * FROM t2 ORDER BY col1; 
+SELECT * FROM t1 UNION SELECT * FROM t2 ORDER BY col1;
-SELECT * FROM t1 UNION (SELECT * FROM t2) ORDER BY col1; 
+# SELECT * FROM t1 UNION (SELECT * FROM t2 UNION SELECT * FROM t3) ORDER BY col1; 
-SELECT * FROM t1 UNION (SELECT * FROM t2 UNION SELECT * FROM t3) ORDER BY col1; 
 # JOIN
 SELECT t1.a, t1.b, t2.c FROM "table" AS t1 JOIN (SELECT * FROM foo JOIN bar ON foo.id = bar.id) t2 ON t1.a = t2.b WHERE (t1.b OR NOT t1.a) AND t2.c = 12.5
 SELECT * FROM t1 JOIN t2 ON c1 = c2;
@@ -15,4 +14,8 @@ SELECT * FROM t1 JOIN t2 ON c1 = c2;
 CREATE TABLE "table" FROM TBL FILE 'students.tbl'
 CREATE TABLE IF NOT EXISTS "table" FROM TBL FILE 'students.tbl'
 # Multiple statements
-CREATE TABLE "table" FROM TBL FILE 'students.tbl'; SELECT * FROM "table";
+CREATE TABLE "table" FROM TBL FILE 'students.tbl'; SELECT * FROM "table";
+# INSERT
+INSERT INTO test_table VALUES (1, 2, 'test');
+INSERT INTO test_table (id, value, name) VALUES (1, 2, 'test');
+INSERT INTO test_table SELECT * FROM students;