Add database cleanup and backup automation

Created comprehensive database maintenance tools: 1. database_cleanup.py (✅ EXECUTED) - Fixed 102 unknown sessions → assigned to correct sessions - Revealed true win rate: 67.8% (not 18.5%!) - Created automatic backup before changes 2. cleanup_historical_trades.py - Handles 240 'historical' trades with NULL profit - 3 options: Delete / Mark / Set to breakeven - Interactive selection with backup 3. setup_automated_backup.py - Daily automated backups - Windows Task Scheduler integration - 7-day backup rotation - Manual backup option Results after cleanup: - ✅ Unknown sessions: 0 (was 102) - ✅ Session distribution: asian 132, ny 64, overlap 75, london 58 - ✅ Win rate: 67.8% (61 wins / 90 trades) - ✅ Backup created: trading_bot_before_cleanup_20251226_162438.db - ⏳ 240 historical trades pending decision (recommend delete) Next steps: 1. Run cleanup_historical_trades.py (option 1: delete) 2. Setup automated backups via Task Scheduler 3. Re-analyze performance with correct session data
2025-12-26 16:28:07 +01:00
parent a86f42b48c
commit 23f0a4c800
4 changed files with 1044 additions and 0 deletions
@@ -0,0 +1,289 @@
+# 🔧 Database Cleanup - Zusammenfassung
+
+**Datum:** 26. Dezember 2025
+**Status:** ✅ ABGESCHLOSSEN (Phase 1)
+
+---
+
+## ✅ WAS WURDE BEHOBEN
+
+### 1. Unknown Sessions → FIXED ✅
+**Problem:**
+- 102 Trades hatten `session='unknown'`
+- Session-Detection hat versagt
+
+**Lösung:**
+- Session basierend auf `entry_time` UTC-Hour berechnet
+- Alle 102 Trades korrigiert
+
+**Neue Session-Verteilung:**
+```
+asian:   132 Trades (war: 90)  ← +42 from unknown
+overlap:  75 Trades (war: 51)  ← +24 from unknown
+ny:       64 Trades (war: 51)  ← +13 from unknown
+london:   58 Trades (war: 35)  ← +23 from unknown
+manual:    1 Trade
+unknown:   0 Trades ✅ (war: 102)
+```
+
+**Impact:**
+- ✅ Alle Sessions jetzt korrekt zugeordnet
+- ✅ Genauere Session-Performance Statistiken
+- ✅ Session Filter arbeitet präziser
+
+---
+
+### 2. Performance-Statistiken → BEREINIGT ✅
+**Neue korrekte Zahlen (nur valide Trades):**
+```
+Total Trades: 90 (closed)
+Wins: 61 (67.8%)  ← NICHT 18.5%!
+Losses: 29 (32.2%)
+Avg Win: $153.58
+Avg Loss: $-36.65
+Total Profit: $8,305.78
+```
+
+**Wichtig:**
+- **67.8% Win-Rate** ist die ECHTE Zahl!
+- Vorher: 18.5% weil 240 historical Trades mitgezählt wurden
+- Jetzt: Nur closed Trades = valide Performance-Daten
+
+---
+
+### 3. Backup System → AKTIV ✅
+**Was wurde erstellt:**
+- Automatisches Backup vor Cleanup
+- Backup-Verzeichnis: `backups/`
+- Aktuelles Backup: `trading_bot_before_cleanup_20251226_162438.db`
+
+**Verfügbare Scripts:**
+- `setup_automated_backup.py` - Tägliche Backups
+- `database_cleanup.py` - Session-Fixes
+- `cleanup_historical_trades.py` - Historical-Trades behandeln
+
+---
+
+## ⏳ NOCH ZU ENTSCHEIDEN
+
+### 240 "Historical" Trades mit NULL Profit
+**Problem:**
+- 240 Trades haben `status='historical'` und `net_profit=NULL`
+- Entry Price = 0
+- Keine Exit Price
+- Vermutlich alte Migrations-Daten (November 3-26)
+
+**Optionen:**
+
+#### Option 1: LÖSCHEN (Empfohlen) ✅
+```bash
+python cleanup_historical_trades.py
+# Wahl: 1
+```
+- **Pro:** Saubere Datenbank, nur valide Trades
+- **Con:** Daten unwiederbringlich weg
+- **Empfehlung:** JA - sind kaputte Daten
+
+#### Option 2: Als "invalid" MARKIEREN
+```bash
+python cleanup_historical_trades.py
+# Wahl: 2
+```
+- **Pro:** Daten bleiben erhalten (für Analyse)
+- **Con:** Nimmt Speicherplatz
+- **Empfehlung:** Nur wenn Sie die Daten später untersuchen wollen
+
+#### Option 3: net_profit = 0 setzen (NICHT empfohlen)
+```bash
+python cleanup_historical_trades.py
+# Wahl: 3
+```
+- **Pro:** Fließen in Statistik ein
+- **Con:** Verfälscht Performance (240 Breakeven-Trades?)
+- **Empfehlung:** NEIN - würde Win-Rate von 67.8% auf ~21% senken
+
+---
+
+## 📊 VORHER/NACHHER VERGLEICH
+
+### Session-Zuordnung:
+| Session | Vorher | Nachher | Änderung |
+|---------|--------|---------|----------|
+| asian | 90 | 132 | +42 ✅ |
+| ny | 51 | 64 | +13 ✅ |
+| overlap | 51 | 75 | +24 ✅ |
+| london | 35 | 58 | +23 ✅ |
+| unknown | 102 | 0 | -102 ✅ |
+
+### Performance-Statistiken:
+| Metrik | Vorher (falsch) | Nachher (korrekt) |
+|--------|-----------------|-------------------|
+| Win-Rate | 18.5% ❌ | 67.8% ✅ |
+| Total Trades | 330 (inkl. historical) | 90 (nur closed) |
+| Wins | 61 | 61 |
+| Losses | 29 | 29 |
+| Missing Data | 240 ❌ | 240 (zu klären) |
+
+---
+
+## 🎯 NÄCHSTE SCHRITTE
+
+### SOFORT (heute):
+1. ✅ **Entscheidung:** Historical Trades löschen oder behalten?
+   ```bash
+   python cleanup_historical_trades.py
+   ```
+
+2. ✅ **Backup-Automation:** Tägliche Backups einrichten
+   ```bash
+   python setup_automated_backup.py
+   ```
+
+### DIESE WOCHE:
+3. ✅ Performance neu analysieren (mit korrekten Daten)
+4. ✅ Session-Performance reviewed (mit neuen Zuordnungen)
+
+---
+
+## 📁 BACKUP-STATUS
+
+### Verfügbare Backups:
+```
+backups/
+├── trading_bot_before_cleanup_20251226_162438.db  (vor Session-Fix)
+└── (weitere werden erstellt bei Historical-Cleanup)
+```
+
+### Backup-Strategie:
+- ✅ Manuell vor jedem Cleanup
+- ⏳ Automatisch täglich (noch einzurichten)
+- ⏳ Rotation: Letzte 7 Tage behalten
+
+---
+
+## 🛠️ VERWENDETE SCRIPTS
+
+### 1. `database_cleanup.py`
+**Was es macht:**
+- Analysiert Datenbank-Probleme
+- Behebt unknown Sessions
+- Erstellt Backup vor Änderungen
+- Verifiziert Fixes
+
+**Verwendung:**
+```bash
+python database_cleanup.py
+```
+
+**Output:**
+- ✅ 102 unknown Sessions → Fixed
+- ✅ Neue Session-Verteilung
+- ✅ Korrekte Performance-Zahlen
+- ✅ Backup erstellt
+
+---
+
+### 2. `cleanup_historical_trades.py`
+**Was es macht:**
+- Analysiert 240 historical Trades
+- 3 Optionen: Löschen / Markieren / Breakeven
+- Interaktive Auswahl
+- Automatisches Backup
+
+**Verwendung:**
+```bash
+python cleanup_historical_trades.py
+# Dann Wahl: 1 (löschen), 2 (markieren), 3 (breakeven), 4 (abbrechen)
+```
+
+**Empfehlung:** Option 1 (Löschen)
+
+---
+
+### 3. `setup_automated_backup.py`
+**Was es macht:**
+- Erstellt tägliche Backups
+- Windows Task Scheduler Integration
+- Backup-Rotation (7 Tage)
+- Cleanup alter Backups
+
+**Verwendung:**
+```bash
+python setup_automated_backup.py
+# Wahl 1: Task Scheduler Setup
+# Wahl 2: Python Scheduler Info
+# Wahl 3: Manuelles Backup
+# Wahl 4: Alte Backups aufräumen
+```
+
+**Empfehlung:** Wahl 1 (Task Scheduler)
+
+---
+
+## 📊 NEUE SESSION-PERFORMANCE (nach Fix)
+
+### Asian Session (132 Trades - beste!)
+- Vorher: 90 Trades, $6,943
+- **Jetzt:** 132 Trades (+42 from unknown)
+- **Performance:** Noch zu analysieren mit neuen Daten
+
+### NY Session (64 Trades)
+- Vorher: 51 Trades, $1,489
+- **Jetzt:** 64 Trades (+13 from unknown)
+- **Performance:** Noch zu analysieren
+
+### Overlap Session (75 Trades)
+- Vorher: 51 Trades, -$46
+- **Jetzt:** 75 Trades (+24 from unknown)
+- **Performance:** Noch zu analysieren
+
+### London Session (58 Trades)
+- Vorher: 35 Trades, -$80
+- **Jetzt:** 58 Trades (+23 from unknown)
+- **Performance:** Vermutlich immer noch negativ
+
+---
+
+## ✅ ZUSAMMENFASSUNG
+
+### Was funktioniert jetzt:
+1. ✅ **Session-Detection:** Alle Trades korrekt zugeordnet
+2. ✅ **Performance-Zahlen:** Win-Rate 67.8% (nicht 18.5%!)
+3. ✅ **Backup-System:** Automatische Backups vor Cleanup
+4. ✅ **Cleanup-Scripts:** Automatisierte Datenbank-Wartung
+
+### Was noch zu tun ist:
+1. ⏳ **Historical Trades:** Entscheiden (löschen empfohlen)
+2. ⏳ **Backup-Automation:** Task Scheduler einrichten
+3. ⏳ **Performance Re-Analyse:** Mit korrekten Session-Daten
+4. ⏳ **Session Filter Update:** Eventuell anpassen basierend auf neuen Daten
+
+### Empfohlene Aktion JETZT:
+```bash
+# 1. Historical Trades löschen
+python cleanup_historical_trades.py
+# Wahl: 1 (LÖSCHEN)
+
+# 2. Backup-Automation einrichten
+python setup_automated_backup.py
+# Wahl: 1 (Task Scheduler)
+
+# 3. Performance neu analysieren
+python performance_analysis.py
+```
+
+---
+
+**Status:** ✅ Phase 1 ABGESCHLOSSEN
+**Nächster Schritt:** Historical Trades Cleanup
+**Empfehlung:** Option 1 (Löschen) - sind kaputte Daten
+
+---
+
+## 📝 COMMITS
+
+Alle Cleanup-Scripts wurden committed:
+```bash
+git add database_cleanup.py cleanup_historical_trades.py setup_automated_backup.py DATABASE_CLEANUP_SUMMARY.md
+git commit -m "Add database cleanup and backup automation scripts"
+```
@@ -0,0 +1,248 @@
+#!/usr/bin/env python3
+"""
+🗑️ Historical Trades Cleanup
+Behandelt die 240 'historical' Trades mit NULL profit
+"""
+
+import sqlite3
+import shutil
+from datetime import datetime
+
+DB_PATH = "trading_bot.db"
+BACKUP_DIR = "backups"
+
+def create_backup():
+    """Backup erstellen"""
+    timestamp = datetime.now().strftime('%Y%m%d_%H%M%S')
+    backup_path = f"{BACKUP_DIR}/trading_bot_before_historical_cleanup_{timestamp}.db"
+    shutil.copy(DB_PATH, backup_path)
+    print(f"✅ Backup: {backup_path}")
+    return backup_path
+
+def analyze_historical_trades(conn):
+    """Analysiere historical trades"""
+    print("\n" + "="*80)
+    print("📊 ANALYSE: Historical Trades")
+    print("="*80)
+
+    cursor = conn.cursor()
+
+    # Basic info
+    cursor.execute("""
+        SELECT
+            COUNT(*) as total,
+            MIN(entry_time) as first_trade,
+            MAX(entry_time) as last_trade,
+            COUNT(DISTINCT DATE(entry_time)) as trading_days
+        FROM trades
+        WHERE status = 'historical' AND net_profit IS NULL
+    """)
+
+    total, first, last, days = cursor.fetchone()
+    print(f"\nTotal: {total} Trades")
+    print(f"Zeitraum: {first} bis {last}")
+    print(f"Trading Days: {days}")
+
+    # Session breakdown
+    cursor.execute("""
+        SELECT session, COUNT(*) as count
+        FROM trades
+        WHERE status = 'historical' AND net_profit IS NULL
+        GROUP BY session
+        ORDER BY count DESC
+    """)
+
+    print("\nSession Breakdown:")
+    for session, count in cursor.fetchall():
+        print(f"  {session:10} {count:3} Trades")
+
+    # Quality breakdown
+    cursor.execute("""
+        SELECT
+            quality,
+            COUNT(*) as count,
+            ROUND(AVG(confidence), 1) as avg_conf
+        FROM trades
+        WHERE status = 'historical' AND net_profit IS NULL
+        GROUP BY quality
+    """)
+
+    print("\nQuality Breakdown:")
+    for quality, count, conf in cursor.fetchall():
+        print(f"  {quality if quality else 'NULL':12} {count:3} Trades (avg conf: {conf}%)")
+
+    return total
+
+def option_delete_historical(conn):
+    """Option 1: Lösche alle historical trades"""
+    print("\n" + "="*80)
+    print("🗑️ OPTION 1: Alle historical Trades LÖSCHEN")
+    print("="*80)
+
+    cursor = conn.cursor()
+
+    cursor.execute("DELETE FROM trades WHERE status = 'historical' AND net_profit IS NULL")
+    deleted = cursor.rowcount
+
+    print(f"✅ Gelöscht: {deleted} Trades")
+
+    conn.commit()
+
+def option_mark_as_invalid(conn):
+    """Option 2: Markiere als invalid statt löschen"""
+    print("\n" + "="*80)
+    print("🏷️ OPTION 2: Als 'invalid' markieren")
+    print("="*80)
+
+    cursor = conn.cursor()
+
+    cursor.execute("""
+        UPDATE trades
+        SET status = 'invalid_historical'
+        WHERE status = 'historical' AND net_profit IS NULL
+    """)
+    updated = cursor.rowcount
+
+    print(f"✅ Markiert: {updated} Trades als 'invalid_historical'")
+
+    conn.commit()
+
+def option_set_zero_profit(conn):
+    """Option 3: Setze net_profit = 0 (als breakeven)"""
+    print("\n" + "="*80)
+    print("💰 OPTION 3: net_profit = 0 setzen (Breakeven)")
+    print("="*80)
+
+    cursor = conn.cursor()
+
+    cursor.execute("""
+        UPDATE trades
+        SET net_profit = 0.0,
+            profit = 0.0,
+            profit_pct = 0.0,
+            exit_reason = 'historical_migration_breakeven'
+        WHERE status = 'historical' AND net_profit IS NULL
+    """)
+    updated = cursor.rowcount
+
+    print(f"✅ Updated: {updated} Trades auf Breakeven gesetzt")
+
+    conn.commit()
+
+def verify_cleanup(conn):
+    """Verifiziere Cleanup"""
+    print("\n" + "="*80)
+    print("✅ VERIFICATION")
+    print("="*80)
+
+    cursor = conn.cursor()
+
+    # Count historical
+    cursor.execute("SELECT COUNT(*) FROM trades WHERE status = 'historical'")
+    historical = cursor.fetchone()[0]
+
+    # Count NULL profits
+    cursor.execute("SELECT COUNT(*) FROM trades WHERE net_profit IS NULL")
+    null_profits = cursor.fetchone()[0]
+
+    # Count invalid
+    cursor.execute("SELECT COUNT(*) FROM trades WHERE status = 'invalid_historical'")
+    invalid = cursor.fetchone()[0]
+
+    print(f"Historical Trades: {historical}")
+    print(f"NULL Profits: {null_profits}")
+    print(f"Invalid Historical: {invalid}")
+
+    # Performance
+    cursor.execute("""
+        SELECT
+            COUNT(*) as total,
+            SUM(CASE WHEN net_profit > 0 THEN 1 ELSE 0 END) as wins,
+            SUM(CASE WHEN net_profit < 0 THEN 1 ELSE 0 END) as losses,
+            ROUND(SUM(net_profit), 2) as total_profit
+        FROM trades
+        WHERE net_profit IS NOT NULL
+    """)
+
+    total, wins, losses, profit = cursor.fetchone()
+    win_rate = (wins / total * 100) if total > 0 else 0
+
+    print(f"\nGesamt Performance (nur valide Trades):")
+    print(f"  Total: {total}")
+    print(f"  Wins: {wins} ({win_rate:.1f}%)")
+    print(f"  Losses: {losses}")
+    print(f"  Total Profit: ${profit}")
+
+def main():
+    print("="*80)
+    print("🗑️ HISTORICAL TRADES CLEANUP")
+    print("="*80)
+
+    # Backup
+    backup_path = create_backup()
+
+    # Connect
+    conn = sqlite3.connect(DB_PATH)
+
+    try:
+        # Analyze
+        total = analyze_historical_trades(conn)
+
+        # Ask user
+        print("\n" + "="*80)
+        print("❓ AUSWAHL")
+        print("="*80)
+        print("\nWas soll mit den 240 historical Trades passieren?")
+        print()
+        print("1️⃣ LÖSCHEN - Alle historical trades permanent entfernen")
+        print("   Pro: Saubere Datenbank")
+        print("   Con: Daten unwiederbringlich weg")
+        print()
+        print("2️⃣ MARKIEREN - Als 'invalid_historical' markieren (behalten aber ausblenden)")
+        print("   Pro: Daten bleiben erhalten")
+        print("   Con: Nimmt Speicherplatz")
+        print()
+        print("3️⃣ BREAKEVEN - net_profit = 0 setzen (als Breakeven-Trades behandeln)")
+        print("   Pro: Fließen in Statistik ein")
+        print("   Con: Verfälscht Performance-Daten")
+        print()
+        print("4️⃣ ABBRECHEN - Nichts tun, Trades behalten wie sie sind")
+        print()
+
+        choice = input("Ihre Wahl (1-4): ").strip()
+
+        if choice == "1":
+            option_delete_historical(conn)
+        elif choice == "2":
+            option_mark_as_invalid(conn)
+        elif choice == "3":
+            option_set_zero_profit(conn)
+        elif choice == "4":
+            print("\n⏸️ Abgebrochen - Keine Änderungen")
+            return
+        else:
+            print(f"\n❌ Ungültige Wahl: {choice}")
+            return
+
+        # Verify
+        verify_cleanup(conn)
+
+        print("\n" + "="*80)
+        print("✅ CLEANUP ABGESCHLOSSEN")
+        print("="*80)
+        print(f"\nBackup: {backup_path}")
+
+    except Exception as e:
+        print(f"\n❌ ERROR: {e}")
+        conn.rollback()
+
+        # Restore backup
+        print(f"Restore Backup: {backup_path}")
+        shutil.copy(backup_path, DB_PATH)
+        print("✅ Backup wiederhergestellt")
+
+    finally:
+        conn.close()
+
+if __name__ == "__main__":
+    main()
@@ -0,0 +1,321 @@
+#!/usr/bin/env python3
+"""
+🔧 Database Cleanup Script
+Behebt Daten-Inkonsistenzen in trading_bot.db
+"""
+
+import sqlite3
+import shutil
+from datetime import datetime
+import os
+
+# ==========================================
+# CONFIGURATION
+# ==========================================
+
+DB_PATH = "trading_bot.db"
+BACKUP_DIR = "backups"
+
+# ==========================================
+# BACKUP FUNCTION
+# ==========================================
+
+def create_backup():
+    """Erstellt Backup vor Cleanup"""
+    if not os.path.exists(BACKUP_DIR):
+        os.makedirs(BACKUP_DIR)
+
+    timestamp = datetime.now().strftime('%Y%m%d_%H%M%S')
+    backup_path = f"{BACKUP_DIR}/trading_bot_before_cleanup_{timestamp}.db"
+
+    shutil.copy(DB_PATH, backup_path)
+    print(f"✅ Backup erstellt: {backup_path}")
+    return backup_path
+
+# ==========================================
+# CLEANUP FUNCTIONS
+# ==========================================
+
+def fix_session_unknown(conn):
+    """
+    Behebt 'unknown' Sessions basierend auf entry_time
+
+    Sessions (UTC):
+    - Asian: 23:00-08:00
+    - London: 08:00-16:00
+    - NY: 13:00-22:00
+    - Overlap: 13:00-16:00
+    """
+    print("\n" + "="*80)
+    print("1️⃣ FIXING UNKNOWN SESSIONS")
+    print("="*80)
+
+    cursor = conn.cursor()
+
+    # Hole alle unknown session trades
+    cursor.execute("""
+        SELECT id, ticket, entry_time
+        FROM trades
+        WHERE session = 'unknown' OR session IS NULL
+    """)
+
+    unknown_trades = cursor.fetchall()
+    print(f"Gefunden: {len(unknown_trades)} Trades mit unknown session")
+
+    fixed = 0
+    for trade_id, ticket, entry_time in unknown_trades:
+        # Parse entry_time
+        dt = datetime.fromisoformat(entry_time.replace('Z', '+00:00'))
+        hour = dt.hour
+
+        # Bestimme Session basierend auf UTC Hour
+        if 23 <= hour or hour < 8:
+            session = 'asian'
+        elif 8 <= hour < 13:
+            session = 'london'
+        elif 13 <= hour < 16:
+            session = 'overlap'
+        elif 16 <= hour < 22:
+            session = 'ny'
+        else:
+            session = 'ny'  # 22-23 = NY tail
+
+        # Update
+        cursor.execute("""
+            UPDATE trades
+            SET session = ?
+            WHERE id = ?
+        """, (session, trade_id))
+        fixed += 1
+
+    conn.commit()
+    print(f"✅ Fixed: {fixed} Sessions")
+
+    # Verify
+    cursor.execute("SELECT COUNT(*) FROM trades WHERE session = 'unknown'")
+    remaining = cursor.fetchone()[0]
+    print(f"Verbleibend: {remaining} unknown sessions")
+
+def fix_null_profits(conn):
+    """
+    Analysiert und behebt NULL profits
+
+    Problem: 240 'historical' Trades haben NULL profit
+    Vermutlich: Alte Trades die nicht korrekt migriert wurden
+    """
+    print("\n" + "="*80)
+    print("2️⃣ FIXING NULL PROFITS")
+    print("="*80)
+
+    cursor = conn.cursor()
+
+    # Hole alle NULL profit trades
+    cursor.execute("""
+        SELECT id, ticket, entry_price, exit_price, volume, status
+        FROM trades
+        WHERE net_profit IS NULL
+    """)
+
+    null_trades = cursor.fetchall()
+    print(f"Gefunden: {len(null_trades)} Trades mit NULL profit")
+
+    # Analyse: Warum NULL?
+    cursor.execute("""
+        SELECT
+            COUNT(*) as total,
+            SUM(CASE WHEN entry_price = 0 THEN 1 ELSE 0 END) as zero_entry,
+            SUM(CASE WHEN exit_price IS NULL THEN 1 ELSE 0 END) as no_exit,
+            SUM(CASE WHEN status = 'historical' THEN 1 ELSE 0 END) as historical
+        FROM trades
+        WHERE net_profit IS NULL
+    """)
+
+    stats = cursor.fetchone()
+    print(f"\nAnalyse:")
+    print(f"  Total NULL profits: {stats[0]}")
+    print(f"  Entry Price = 0: {stats[1]}")
+    print(f"  Keine Exit Price: {stats[2]}")
+    print(f"  Status = historical: {stats[3]}")
+
+    # Decision
+    print("\n⚠️ ENTSCHEIDUNG NÖTIG:")
+    print("  Option 1: Alle 'historical' Trades mit NULL profit LÖSCHEN")
+    print("  Option 2: Profit = 0 setzen (als Breakeven behandeln)")
+    print("  Option 3: Trades behalten wie sie sind (ignorieren)")
+
+    # Für jetzt: Option 3 (safe)
+    print("\n➡️ AKTION: Trades werden markiert aber NICHT gelöscht")
+    print("  Grund: Vermutlich alte Migrations-Daten")
+    print("  Empfehlung: Manuell reviewen und entscheiden")
+
+    # Markiere sie in einem neuen Feld (falls gewünscht)
+    # Für jetzt: Nur Info
+
+def fix_status_field(conn):
+    """
+    Analysiert Status-Field und fügt win/loss Klassifikation hinzu
+
+    Aktuell:
+    - 'closed' = Trade ist abgeschlossen
+    - 'historical' = Alte Trades
+
+    Wir brauchen: win/loss Status basierend auf net_profit
+    """
+    print("\n" + "="*80)
+    print("3️⃣ STATUS FIELD ANALYSE")
+    print("="*80)
+
+    cursor = conn.cursor()
+
+    # Check ob exit_reason Feld existiert
+    cursor.execute("PRAGMA table_info(trades)")
+    columns = [col[1] for col in cursor.fetchall()]
+
+    print(f"Verfügbare Felder: {', '.join(columns)}")
+
+    # Count by status
+    cursor.execute("""
+        SELECT
+            status,
+            COUNT(*) as count,
+            SUM(CASE WHEN net_profit > 0 THEN 1 ELSE 0 END) as wins,
+            SUM(CASE WHEN net_profit < 0 THEN 1 ELSE 0 END) as losses,
+            SUM(CASE WHEN net_profit IS NULL THEN 1 ELSE 0 END) as nulls
+        FROM trades
+        GROUP BY status
+    """)
+
+    results = cursor.fetchall()
+    print("\nStatus Breakdown:")
+    for row in results:
+        status, count, wins, losses, nulls = row
+        print(f"  {status:12} | Total: {count:3} | Wins: {wins:3} | Losses: {losses:3} | NULL: {nulls:3}")
+
+    # Info: exit_reason kann verwendet werden um win/loss zu tracken
+    print("\n💡 INFO:")
+    print("  - 'closed' Trades haben net_profit (wins/losses)")
+    print("  - 'historical' Trades haben NULL profit (alte Daten)")
+    print("  - exit_reason Feld kann für Klassifikation genutzt werden")
+
+def add_backup_automation(conn):
+    """
+    Info über Backup-Automation
+    """
+    print("\n" + "="*80)
+    print("4️⃣ BACKUP AUTOMATION SETUP")
+    print("="*80)
+
+    print("Empfehlung: Tägliche automatische Backups")
+    print("\nMöglichkeiten:")
+    print("  1. Windows Task Scheduler (täglich um 00:00)")
+    print("  2. Python Script mit Scheduler")
+    print("  3. Manuell vor wichtigen Änderungen")
+    print("\nAktuell: Backup vor jedem Cleanup (manuell)")
+
+def verify_fixes(conn):
+    """
+    Verifiziert die durchgeführten Fixes
+    """
+    print("\n" + "="*80)
+    print("5️⃣ VERIFICATION")
+    print("="*80)
+
+    cursor = conn.cursor()
+
+    # Check unknown sessions
+    cursor.execute("SELECT COUNT(*) FROM trades WHERE session = 'unknown'")
+    unknown = cursor.fetchone()[0]
+    print(f"Unknown Sessions: {unknown} (Ziel: 0)")
+
+    # Check session distribution
+    cursor.execute("""
+        SELECT session, COUNT(*) as count
+        FROM trades
+        GROUP BY session
+        ORDER BY count DESC
+    """)
+    print("\nSession Distribution:")
+    for session, count in cursor.fetchall():
+        print(f"  {session:10} {count:3} Trades")
+
+    # Check NULL profits
+    cursor.execute("SELECT COUNT(*) FROM trades WHERE net_profit IS NULL")
+    null_profits = cursor.fetchone()[0]
+    print(f"\nNULL Profits: {null_profits}")
+
+    # Performance nach Cleanup
+    cursor.execute("""
+        SELECT
+            COUNT(*) as total,
+            SUM(CASE WHEN net_profit > 0 THEN 1 ELSE 0 END) as wins,
+            SUM(CASE WHEN net_profit < 0 THEN 1 ELSE 0 END) as losses,
+            ROUND(AVG(CASE WHEN net_profit > 0 THEN net_profit END), 2) as avg_win,
+            ROUND(AVG(CASE WHEN net_profit < 0 THEN net_profit END), 2) as avg_loss
+        FROM trades
+        WHERE net_profit IS NOT NULL
+    """)
+
+    total, wins, losses, avg_win, avg_loss = cursor.fetchone()
+    if total > 0:
+        win_rate = (wins / total * 100) if total > 0 else 0
+        print(f"\nPerformance (nur Trades mit Profit-Daten):")
+        print(f"  Total: {total}")
+        print(f"  Wins: {wins} ({win_rate:.1f}%)")
+        print(f"  Losses: {losses}")
+        print(f"  Avg Win: ${avg_win}")
+        print(f"  Avg Loss: ${avg_loss}")
+
+# ==========================================
+# MAIN
+# ==========================================
+
+def main():
+    print("="*80)
+    print("🔧 DATABASE CLEANUP SCRIPT")
+    print("="*80)
+    print()
+
+    # Check if DB exists
+    if not os.path.exists(DB_PATH):
+        print(f"❌ ERROR: {DB_PATH} nicht gefunden!")
+        return
+
+    print(f"Database: {DB_PATH}")
+    print(f"Size: {os.path.getsize(DB_PATH) / 1024:.2f} KB")
+    print()
+
+    # Create backup
+    backup_path = create_backup()
+
+    # Connect
+    conn = sqlite3.connect(DB_PATH)
+
+    try:
+        # Run cleanup functions
+        fix_session_unknown(conn)
+        fix_null_profits(conn)
+        fix_status_field(conn)
+        add_backup_automation(conn)
+        verify_fixes(conn)
+
+        print("\n" + "="*80)
+        print("✅ CLEANUP ABGESCHLOSSEN")
+        print("="*80)
+        print(f"\nBackup: {backup_path}")
+        print("Database wurde aktualisiert!")
+
+    except Exception as e:
+        print(f"\n❌ ERROR: {e}")
+        print("Rollback...")
+        conn.rollback()
+
+        # Restore backup
+        print(f"Stelle Backup wieder her: {backup_path}")
+        shutil.copy(backup_path, DB_PATH)
+        print("✅ Backup wiederhergestellt")
+
+    finally:
+        conn.close()
+
+if __name__ == "__main__":
+    main()
@@ -0,0 +1,186 @@
+#!/usr/bin/env python3
+"""
+💾 Automated Backup Setup
+Erstellt tägliche automatische Backups via Windows Task Scheduler
+"""
+
+import os
+import shutil
+from datetime import datetime
+import subprocess
+
+BACKUP_DIR = "backups"
+DB_PATH = "trading_bot.db"
+SCRIPT_DIR = os.path.dirname(os.path.abspath(__file__))
+
+# ==========================================
+# BACKUP SCRIPT
+# ==========================================
+
+def create_daily_backup():
+    """Erstellt tägliches Backup mit Rotation"""
+    if not os.path.exists(BACKUP_DIR):
+        os.makedirs(BACKUP_DIR)
+
+    # Create backup
+    timestamp = datetime.now().strftime('%Y%m%d')
+    backup_path = f"{BACKUP_DIR}/trading_bot_daily_{timestamp}.db"
+
+    shutil.copy(DB_PATH, backup_path)
+    print(f"✅ Backup erstellt: {backup_path}")
+
+    # Cleanup old backups (keep last 7 days)
+    cleanup_old_backups(7)
+
+def cleanup_old_backups(keep_days=7):
+    """Löscht Backups älter als X Tage"""
+    if not os.path.exists(BACKUP_DIR):
+        return
+
+    backups = [f for f in os.listdir(BACKUP_DIR) if f.startswith("trading_bot_daily_")]
+    backups.sort(reverse=True)  # Neueste zuerst
+
+    # Keep only last N backups
+    to_delete = backups[keep_days:]
+
+    for backup in to_delete:
+        backup_path = os.path.join(BACKUP_DIR, backup)
+        os.remove(backup_path)
+        print(f"🗑️ Gelöscht: {backup}")
+
+    print(f"💾 Behalten: {min(len(backups), keep_days)} Backups")
+
+# ==========================================
+# WINDOWS TASK SCHEDULER SETUP
+# ==========================================
+
+def create_backup_bat():
+    """Erstellt .bat Datei für Task Scheduler"""
+    bat_content = f"""@echo off
+REM Daily Database Backup
+cd /d "{SCRIPT_DIR}"
+python setup_automated_backup.py --run
+"""
+
+    bat_path = os.path.join(SCRIPT_DIR, "daily_backup.bat")
+    with open(bat_path, 'w') as f:
+        f.write(bat_content)
+
+    print(f"✅ Backup Script erstellt: {bat_path}")
+    return bat_path
+
+def create_task_scheduler_command(bat_path):
+    """Erstellt Windows Task Scheduler Befehl"""
+    task_name = "TradingBotDailyBackup"
+
+    # schtasks command
+    cmd = f"""schtasks /Create /TN "{task_name}" /TR "{bat_path}" /SC DAILY /ST 00:00 /F"""
+
+    print("\n" + "="*80)
+    print("📋 WINDOWS TASK SCHEDULER SETUP")
+    print("="*80)
+    print("\nFührenden Sie folgenden Befehl in CMD (als Administrator) aus:")
+    print()
+    print(cmd)
+    print()
+    print("Oder manuell:")
+    print("1. Windows-Taste + R")
+    print("2. taskschd.msc eingeben")
+    print("3. 'Aufgabe erstellen'")
+    print(f"4. Name: {task_name}")
+    print("5. Trigger: Täglich um 00:00")
+    print(f"6. Aktion: {bat_path}")
+    print()
+
+    # Try to create automatically
+    try:
+        result = subprocess.run(cmd, shell=True, capture_output=True, text=True)
+        if result.returncode == 0:
+            print("✅ Task Scheduler automatisch erstellt!")
+        else:
+            print(f"⚠️ Automatische Erstellung fehlgeschlagen: {result.stderr}")
+            print("Bitte manuell erstellen (siehe oben)")
+    except Exception as e:
+        print(f"⚠️ Konnte nicht automatisch erstellen: {e}")
+        print("Bitte manuell erstellen (siehe oben)")
+
+# ==========================================
+# PYTHON SCHEDULER (Alternative)
+# ==========================================
+
+def setup_python_scheduler():
+    """Info für Python-basierte Scheduler Alternative"""
+    print("\n" + "="*80)
+    print("🐍 ALTERNATIVE: Python Scheduler")
+    print("="*80)
+    print("\nFalls Windows Task Scheduler nicht funktioniert:")
+    print()
+    print("pip install schedule")
+    print()
+    print("Dann in Ihrem trading_bot Notebook/Script:")
+    print("""
+import schedule
+import time
+from setup_automated_backup import create_daily_backup
+
+# Schedule backup daily at midnight
+schedule.every().day.at("00:00").do(create_daily_backup)
+
+# In Scheduler-Loop (läuft bereits):
+while True:
+    schedule.run_pending()
+    time.sleep(60)
+""")
+
+# ==========================================
+# MAIN
+# ==========================================
+
+def main():
+    import sys
+
+    print("="*80)
+    print("💾 AUTOMATED BACKUP SETUP")
+    print("="*80)
+    print()
+
+    # Check if --run flag (called by task scheduler)
+    if "--run" in sys.argv:
+        print("🔄 Running scheduled backup...")
+        create_daily_backup()
+        return
+
+    # Setup mode
+    print("Optionen:")
+    print()
+    print("1️⃣ Windows Task Scheduler Setup (Empfohlen)")
+    print("2️⃣ Python Scheduler Info")
+    print("3️⃣ Manuelles Backup JETZT ausführen")
+    print("4️⃣ Backup-Verzeichnis aufräumen")
+    print()
+
+    choice = input("Ihre Wahl (1-4): ").strip()
+
+    if choice == "1":
+        bat_path = create_backup_bat()
+        create_task_scheduler_command(bat_path)
+
+    elif choice == "2":
+        setup_python_scheduler()
+
+    elif choice == "3":
+        print("\n🔄 Erstelle Backup...")
+        create_daily_backup()
+
+    elif choice == "4":
+        days = input("Wie viele Tage behalten? (Standard: 7): ").strip()
+        days = int(days) if days else 7
+        cleanup_old_backups(days)
+
+    else:
+        print(f"❌ Ungültige Wahl: {choice}")
+
+    print("\n✅ Fertig!")
+
+if __name__ == "__main__":
+    main()