Merge pull request #1156 from VXNCXNX/fix/leading-quoted-field-delimiter

fix: a delimiter inside a quoted field at line start splits the column

Author: noborus

oviewer/utils.go

@@ -65,7 +65,40 @@ func allIndex(s string, substr string, reg *regexp.Regexp) [][]int {
 	if reg != nil {
 		return reg.FindAllStringIndex(s, -1)
 	}
-	return allStringIndex(s, substr)
+	return allDelimiterIndex(s, substr)
+}
+
+// allDelimiterIndex returns all delimiter positions.
+// A delimiter inside a double-quoted field is part of the field, not a
+// delimiter, including when the quoted field is the first one on the line.
+func allDelimiterIndex(s string, substr string) [][]int {
+	if len(substr) == 0 {
+		return nil
+	}
+	var result [][]int
+	width := len(substr)
+	offSet := 0
+	if strings.HasPrefix(s, `"`) {
+		s, offSet = skipQuoted(s, offSet)
+	}
+	for pos := strings.Index(s, substr); pos != -1; pos = strings.Index(s, substr) {
+		s = s[pos+width:]
+		result = append(result, []int{pos + offSet, pos + offSet + width})
+		offSet += pos + width
+
+		if strings.HasPrefix(s, `"`) {
+			s, offSet = skipQuoted(s, offSet)
+		}
+	}
+	return result
+}
+
+// skipQuoted skips the double-quoted field at the beginning of s and returns
+// the remainder with the updated offset. An unclosed quote skips only the
+// opening quote, which is what the delimiter search did before.
+func skipQuoted(s string, offSet int) (string, int) {
+	qpos := strings.Index(s[1:], `"`)
+	return s[qpos+2:], offSet + qpos + 2
 }
 
 // allStringIndex returns all matching string positions.

oviewer/utils_test.go

@@ -513,6 +513,106 @@ func Test_allStringIndex(t *testing.T) {
 	}
 }
 
+func Test_allDelimiterIndex(t *testing.T) {
+	t.Parallel()
+	type args struct {
+		s      string
+		substr string
+	}
+	tests := []struct {
+		name string
+		args args
+		want [][]int
+	}{
+		{
+			name: "test1",
+			args: args{
+				s:      "a,b,c",
+				substr: ",",
+			},
+			want: [][]int{
+				{1, 2},
+				{3, 4},
+			},
+		},
+		{
+			name: "testNone",
+			args: args{
+				s:      "a,b,c",
+				substr: "@",
+			},
+			want: nil,
+		},
+		{
+			name: "testNoSubstr",
+			args: args{
+				s:      "a,b,c",
+				substr: "",
+			},
+			want: nil,
+		},
+		{
+			name: "testDoubleQuote",
+			args: args{
+				s:      `a,"b,c",d`,
+				substr: ",",
+			},
+			want: [][]int{
+				{1, 2},
+				{7, 8},
+			},
+		},
+		{
+			name: "testLeadingDoubleQuote",
+			args: args{
+				s:      `"a,b",c,d`,
+				substr: ",",
+			},
+			want: [][]int{
+				{5, 6},
+				{7, 8},
+			},
+		},
+		{
+			name: "testLeadingDoubleQuoteOnly",
+			args: args{
+				s:      `"a,b"`,
+				substr: ",",
+			},
+			want: nil,
+		},
+		{
+			name: "testLeadingEmptyDoubleQuote",
+			args: args{
+				s:      `"",a,b`,
+				substr: ",",
+			},
+			want: [][]int{
+				{2, 3},
+				{4, 5},
+			},
+		},
+		{
+			name: "testLeadingUnclosedDoubleQuote",
+			args: args{
+				s:      `"a,b`,
+				substr: ",",
+			},
+			want: [][]int{
+				{2, 3},
+			},
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			t.Parallel()
+			if got := allDelimiterIndex(tt.args.s, tt.args.substr); !reflect.DeepEqual(got, tt.want) {
+				t.Errorf("allDelimiterIndex() = %v, want %v", got, tt.want)
+			}
+		})
+	}
+}
+
 func Test_abs(t *testing.T) {
 	t.Parallel()
 	tests := []struct {