QT qDeleteAll and new afterwards same memory address and crash

Question

I have a weird behavior when using qDeleteAll on a QMap<QString, Record*>. First I delete my values in the QMap with qDeleteAll and then I use m_data.clear() to empty the QMap.

void ClassName::function1(QMap<QString, QString> filter)
{
    qDebug() << "function1 called";
    QString query = "postgrestAdress?";
    if (filter.size()){
        for (QMap<QString, QString>::iterator i = filter.begin(); i != filter.end(); i++ ) {
            query.append("&").append(i.key()).append("=").append(i.value());
        }
    }

    postgrestLambdaFucntion(query,[this, query](QList<QMap<QString,QString>> result)
    {
        qDeleteAll(m_data);

        m_data.clear();
        m_users.clear();

        for(auto &line : result)
        {
            Record *record = Mapper::mapQMapToRecord(line);
            qDebug() << "created Record:" << record << " with id:" << record->id;
            m_data.insert(line.value("material_number")+ "_" +QString::number(record->id), record);
            if (!m_users.contains(line.value("user"))){
                m_users.append(line.value("user"));
            }
        }
        nextFunction(0);
    });
}

Consider this as a record:

class Record : public QObject
{
    Q_OBJECT

public:
    explicit Record(QObject *parent = nullptr) : QObject(parent){};
    ~Record() {
        qDebug() << "record " << this << "with id" << this->id << " is destroyed";
    };

    int id = 0;
    QString material_number = "";
    QString user = "";
} 

When I call function1 with an already populated m_data list I get this debug information:

>Returing table from postgrest...
>record  Record(0x564813686040) with id 1  is destroyed
>record  Record(0x5648136f74b0) with id 33  is destroyed
>record  Record(0x56481376a6b0) with id 15  is destroyed
>record  Record(0x564813708c00) with id 43  is destroyed
>record  Record(0x564813c28090) with id 44  is destroyed
>record  Record(0x564813768eb0) with id 2  is destroyed
>record  Record(0x5648136eb7a0) with id 14  is destroyed
>created Record: Record(0x564813768eb0)  with id: 15
>created Record: Record(0x564813c28090)  with id: 2
>created Record: Record(0x564813708c00)  with id: 43
>created Record: Record(0x56481376a6b0)  with id: 33
>created Record: Record(0x5648136f74b0)  with id: 14
>created Record: Record(0x564813686040)  with id: 1
>created Record: Record(0x564813de45b0)  with id: 44

Inside the Mapper-Function I create new Record-Classes

Record* Mapper::mapQMapToRecord(QMap<QString, QString> line)
{

    Record* record = new Record(nullptr);
    record->id = line.value("id").toInt();
    record->user = line.value("user");
    return record;

}

Note that the Memory-Address of those new created objects are the same as the ones qDeleteAll removed. In the next function I make more requests and add data to the records, which are now "not available" and I get nullptr exceptions.

If I don't use "qDeleteAll" it works. I can also use a for-loop and use "deleteLater" from QObject without errors. I just want to know this behaviour better or why it happens. (it happens in debug and release builds)

Anybody has insights for me?

Best Regards

exa.byte