fix splitting multiple SQLs

2025-09-20 21:57:56 +08:00 · 2025-09-03 17:47:24 -07:00
parent ea758d0d9f
commit 1db1206827
2 changed files with 253 additions and 23 deletions
--- a/weed/command/sql_test.go
+++ b/weed/command/sql_test.go
@@ -0,0 +1,149 @@
+package command
+
+import (
+	"reflect"
+	"testing"
+)
+
+func TestSplitSQLStatements(t *testing.T) {
+	tests := []struct {
+		name     string
+		input    string
+		expected []string
+	}{
+		{
+			name:     "Simple single statement",
+			input:    "SELECT * FROM users",
+			expected: []string{"SELECT * FROM users"},
+		},
+		{
+			name:     "Multiple statements",
+			input:    "SELECT * FROM users; SELECT * FROM orders;",
+			expected: []string{"SELECT * FROM users", "SELECT * FROM orders"},
+		},
+		{
+			name:     "Semicolon in single quotes",
+			input:    "SELECT 'hello;world' FROM users; SELECT * FROM orders;",
+			expected: []string{"SELECT 'hello;world' FROM users", "SELECT * FROM orders"},
+		},
+		{
+			name:     "Semicolon in double quotes",
+			input:    `SELECT "column;name" FROM users; SELECT * FROM orders;`,
+			expected: []string{`SELECT "column;name" FROM users`, "SELECT * FROM orders"},
+		},
+		{
+			name:     "Escaped quotes in strings",
+			input:    `SELECT 'don''t split; here' FROM users; SELECT * FROM orders;`,
+			expected: []string{`SELECT 'don''t split; here' FROM users`, "SELECT * FROM orders"},
+		},
+		{
+			name:     "Escaped quotes in identifiers",
+			input:    `SELECT "column""name" FROM users; SELECT * FROM orders;`,
+			expected: []string{`SELECT "column""name" FROM users`, "SELECT * FROM orders"},
+		},
+		{
+			name:     "Single line comment",
+			input:    "SELECT * FROM users; -- This is a comment\nSELECT * FROM orders;",
+			expected: []string{"SELECT * FROM users", "-- This is a comment\nSELECT * FROM orders"},
+		},
+		{
+			name:     "Single line comment with semicolon",
+			input:    "SELECT * FROM users; -- Comment with; semicolon\nSELECT * FROM orders;",
+			expected: []string{"SELECT * FROM users", "-- Comment with; semicolon\nSELECT * FROM orders"},
+		},
+		{
+			name:     "Multi-line comment",
+			input:    "SELECT * FROM users; /* Multi-line\ncomment */ SELECT * FROM orders;",
+			expected: []string{"SELECT * FROM users", "/* Multi-line\ncomment */ SELECT * FROM orders"},
+		},
+		{
+			name:     "Multi-line comment with semicolon",
+			input:    "SELECT * FROM users; /* Comment with; semicolon */ SELECT * FROM orders;",
+			expected: []string{"SELECT * FROM users", "/* Comment with; semicolon */ SELECT * FROM orders"},
+		},
+		{
+			name:     "Complex mixed case",
+			input:    `SELECT 'test;string', "quoted;id" FROM users; -- Comment; here
+			/* Another; comment */ 
+			INSERT INTO users VALUES ('name''s value', "id""field");`,
+			expected: []string{
+				`SELECT 'test;string', "quoted;id" FROM users`,
+				`-- Comment; here
+			/* Another; comment */ 
+			INSERT INTO users VALUES ('name''s value', "id""field")`,
+			},
+		},
+		{
+			name:     "Empty statements filtered",
+			input:    "SELECT * FROM users;;; SELECT * FROM orders;",
+			expected: []string{"SELECT * FROM users", "SELECT * FROM orders"},
+		},
+		{
+			name:     "Whitespace handling",
+			input:    "  SELECT * FROM users  ;  SELECT * FROM orders  ;  ",
+			expected: []string{"SELECT * FROM users", "SELECT * FROM orders"},
+		},
+		{
+			name:     "Single statement without semicolon",
+			input:    "SELECT * FROM users",
+			expected: []string{"SELECT * FROM users"},
+		},
+		{
+			name:     "Empty query",
+			input:    "",
+			expected: []string{},
+		},
+		{
+			name:     "Only whitespace",
+			input:    "   \n\t   ",
+			expected: []string{},
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			result := splitSQLStatements(tt.input)
+			if !reflect.DeepEqual(result, tt.expected) {
+				t.Errorf("splitSQLStatements() = %v, expected %v", result, tt.expected)
+			}
+		})
+	}
+}
+
+func TestSplitSQLStatements_EdgeCases(t *testing.T) {
+	tests := []struct {
+		name     string
+		input    string
+		expected []string
+	}{
+		{
+			name:     "Nested comments are not supported but handled gracefully",
+			input:    "SELECT * FROM users; /* Outer /* inner */ comment */ SELECT * FROM orders;",
+			expected: []string{"SELECT * FROM users", "/* Outer /* inner */ comment */ SELECT * FROM orders"},
+		},
+		{
+			name:     "Unterminated string (malformed SQL)",
+			input:    "SELECT 'unterminated string; SELECT * FROM orders;",
+			expected: []string{"SELECT 'unterminated string; SELECT * FROM orders;"},
+		},
+		{
+			name:     "Unterminated comment (malformed SQL)",
+			input:    "SELECT * FROM users; /* unterminated comment",
+			expected: []string{"SELECT * FROM users", "/* unterminated comment"},
+		},
+		{
+			name:     "Multiple semicolons in quotes",
+			input:    "SELECT ';;;' FROM users; SELECT ';;;' FROM orders;",
+			expected: []string{"SELECT ';;;' FROM users", "SELECT ';;;' FROM orders"},
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			result := splitSQLStatements(tt.input)
+			if !reflect.DeepEqual(result, tt.expected) {
+				t.Errorf("splitSQLStatements() = %v, expected %v", result, tt.expected)
+			}
+		})
+	}
+}