Ihre Datenvorlage für das Mitarbeiter-Onboarding

/*
This SQL-like pseudo-code represents the logic to be built across multiple Integration Center jobs.
The final event log is created by combining the output of these logical blocks.
Replace '{StartDate}' and '{EndDate}' with the desired date range for the extraction.
*/

-- Block 1: From the ONB2Process based Integration
SELECT
  p.processId AS OnboardingCaseId,
  'Onboarding Process Initiated' AS ActivityName,
  p.createdDate AS EventTimestamp,
  'SAP SuccessFactors' AS SourceSystem,
  GETDATE() AS LastDataUpdateTimestamp,
  p.createdDate AS EndTime,
  u.userId AS NewHireEmployeeId,
  ej.department AS Department,
  ej.position AS PositionTitle,
  ej.managerId AS HiringManager,
  p.processor AS OnboardingCoordinator,
  p.targetDate AS OfficialStartDate
FROM ONB2Process AS p
LEFT JOIN User AS u ON p.processTrigger.userId = u.userId
LEFT JOIN EmpJob AS ej ON u.userId = ej.userId AND ej.isPrimary = 'true'
WHERE p.createdDate BETWEEN '{StartDate}' AND '{EndDate}'

UNION ALL

-- Block 2: From the ONB2Process based Integration (Calculated Event)
SELECT
  p.processId AS OnboardingCaseId,
  'First Day Occurred' AS ActivityName,
  p.targetDate AS EventTimestamp,
  'SAP SuccessFactors' AS SourceSystem,
  GETDATE() AS LastDataUpdateTimestamp,
  p.targetDate AS EndTime,
  u.userId AS NewHireEmployeeId,
  ej.department AS Department,
  ej.position AS PositionTitle,
  ej.managerId AS HiringManager,
  p.processor AS OnboardingCoordinator,
  p.targetDate AS OfficialStartDate
FROM ONB2Process AS p
LEFT JOIN User AS u ON p.processTrigger.userId = u.userId
LEFT JOIN EmpJob AS ej ON u.userId = ej.userId AND ej.isPrimary = 'true'
WHERE p.targetDate BETWEEN '{StartDate}' AND '{EndDate}'

UNION ALL

-- Block 3: From the ONB2ProcessTask based Integration (multiple mappings)
SELECT
  t.process.processId AS OnboardingCaseId,
  CASE 
    WHEN t.taskName = '[Task name for Paperwork Sent]' THEN 'Pre-Day 1 Paperwork Sent'
    WHEN t.taskName = '[Task name for Background Check]' AND t.taskStatus = 'STARTED' THEN 'Background Check Initiated'
    WHEN t.taskName = '[Task name for Background Check]' AND t.taskStatus = 'COMPLETED' THEN 'Background Check Completed'
    WHEN t.taskName = '[Task name for Paperwork Completion]' THEN 'Pre-Day 1 Paperwork Completed'
    WHEN t.taskName = '[Task name for HR Profile Creation]' THEN 'HR Profile Created'
    WHEN t.taskName = '[Task name for Manager Tasks]' AND t.taskStatus = 'STARTED' THEN 'Manager Tasks Assigned'
    WHEN t.taskName = '[Task name for IT Request]' THEN 'IT Provisioning Requested'
    WHEN t.taskName = '[Task name for Equipment Order]' THEN 'Equipment Ordered'
    WHEN t.taskName = '[Task name for Equipment Delivery]' THEN 'Equipment Delivered'
    WHEN t.taskName = '[Task name for Training Assignment]' THEN 'Mandatory Training Assigned'
    WHEN t.taskName = '[Task name for Training Completion]' THEN 'Mandatory Training Completed'
    WHEN t.taskName = '[Task name for 30-Day Check-in]' THEN '30-Day Check-in Performed'
  END AS ActivityName,
  t.completedDate AS EventTimestamp, -- Or t.createdDate depending on the activity
  'SAP SuccessFactors' AS SourceSystem,
  GETDATE() AS LastDataUpdateTimestamp,
  t.completedDate AS EndTime,
  t.process.processTrigger.userId AS NewHireEmployeeId,
  ej.department AS Department,
  ej.position AS PositionTitle,
  ej.managerId AS HiringManager,
  t.process.processor AS OnboardingCoordinator,
  t.process.targetDate AS OfficialStartDate
FROM ONB2ProcessTask AS t
LEFT JOIN User AS u ON t.process.processTrigger.userId = u.userId
LEFT JOIN EmpJob AS ej ON u.userId = ej.userId AND ej.isPrimary = 'true'
WHERE t.lastModifiedDate BETWEEN '{StartDate}' AND '{EndDate}' AND ActivityName IS NOT NULL

UNION ALL

-- Block 4: From the ONB2Process based Integration
SELECT
  p.processId AS OnboardingCaseId,
  'Onboarding Process Completed' AS ActivityName,
  p.lastModifiedDate AS EventTimestamp, -- Assuming last modification on a completed status marks the end
  'SAP SuccessFactors' AS SourceSystem,
  GETDATE() AS LastDataUpdateTimestamp,
  p.lastModifiedDate AS EndTime,
  u.userId AS NewHireEmployeeId,
  ej.department AS Department,
  ej.position AS PositionTitle,
  ej.managerId AS HiringManager,
  p.processor AS OnboardingCoordinator,
  p.targetDate AS OfficialStartDate
FROM ONB2Process AS p
LEFT JOIN User AS u ON p.processTrigger.userId = u.userId
LEFT JOIN EmpJob AS ej ON u.userId = ej.userId AND ej.isPrimary = 'true'
WHERE p.processStatus = 'COMPLETED' AND p.lastModifiedDate BETWEEN '{StartDate}' AND '{EndDate}';

/*
This SQL-like pseudo-code represents the logic to be built across multiple Integration Center jobs.
The final event log is created by combining the output of these logical blocks.
Replace '{StartDate}' and '{EndDate}' with the desired date range for the extraction.
*/

-- Block 1: From the ONB2Process based Integration
SELECT
  p.processId AS OnboardingCaseId,
  'Onboarding Process Initiated' AS ActivityName,
  p.createdDate AS EventTimestamp,
  'SAP SuccessFactors' AS SourceSystem,
  GETDATE() AS LastDataUpdateTimestamp,
  p.createdDate AS EndTime,
  u.userId AS NewHireEmployeeId,
  ej.department AS Department,
  ej.position AS PositionTitle,
  ej.managerId AS HiringManager,
  p.processor AS OnboardingCoordinator,
  p.targetDate AS OfficialStartDate
FROM ONB2Process AS p
LEFT JOIN User AS u ON p.processTrigger.userId = u.userId
LEFT JOIN EmpJob AS ej ON u.userId = ej.userId AND ej.isPrimary = 'true'
WHERE p.createdDate BETWEEN '{StartDate}' AND '{EndDate}'

UNION ALL

-- Block 2: From the ONB2Process based Integration (Calculated Event)
SELECT
  p.processId AS OnboardingCaseId,
  'First Day Occurred' AS ActivityName,
  p.targetDate AS EventTimestamp,
  'SAP SuccessFactors' AS SourceSystem,
  GETDATE() AS LastDataUpdateTimestamp,
  p.targetDate AS EndTime,
  u.userId AS NewHireEmployeeId,
  ej.department AS Department,
  ej.position AS PositionTitle,
  ej.managerId AS HiringManager,
  p.processor AS OnboardingCoordinator,
  p.targetDate AS OfficialStartDate
FROM ONB2Process AS p
LEFT JOIN User AS u ON p.processTrigger.userId = u.userId
LEFT JOIN EmpJob AS ej ON u.userId = ej.userId AND ej.isPrimary = 'true'
WHERE p.targetDate BETWEEN '{StartDate}' AND '{EndDate}'

UNION ALL

-- Block 3: From the ONB2ProcessTask based Integration (multiple mappings)
SELECT
  t.process.processId AS OnboardingCaseId,
  CASE 
    WHEN t.taskName = '[Task name for Paperwork Sent]' THEN 'Pre-Day 1 Paperwork Sent'
    WHEN t.taskName = '[Task name for Background Check]' AND t.taskStatus = 'STARTED' THEN 'Background Check Initiated'
    WHEN t.taskName = '[Task name for Background Check]' AND t.taskStatus = 'COMPLETED' THEN 'Background Check Completed'
    WHEN t.taskName = '[Task name for Paperwork Completion]' THEN 'Pre-Day 1 Paperwork Completed'
    WHEN t.taskName = '[Task name for HR Profile Creation]' THEN 'HR Profile Created'
    WHEN t.taskName = '[Task name for Manager Tasks]' AND t.taskStatus = 'STARTED' THEN 'Manager Tasks Assigned'
    WHEN t.taskName = '[Task name for IT Request]' THEN 'IT Provisioning Requested'
    WHEN t.taskName = '[Task name for Equipment Order]' THEN 'Equipment Ordered'
    WHEN t.taskName = '[Task name for Equipment Delivery]' THEN 'Equipment Delivered'
    WHEN t.taskName = '[Task name for Training Assignment]' THEN 'Mandatory Training Assigned'
    WHEN t.taskName = '[Task name for Training Completion]' THEN 'Mandatory Training Completed'
    WHEN t.taskName = '[Task name for 30-Day Check-in]' THEN '30-Day Check-in Performed'
  END AS ActivityName,
  t.completedDate AS EventTimestamp, -- Or t.createdDate depending on the activity
  'SAP SuccessFactors' AS SourceSystem,
  GETDATE() AS LastDataUpdateTimestamp,
  t.completedDate AS EndTime,
  t.process.processTrigger.userId AS NewHireEmployeeId,
  ej.department AS Department,
  ej.position AS PositionTitle,
  ej.managerId AS HiringManager,
  t.process.processor AS OnboardingCoordinator,
  t.process.targetDate AS OfficialStartDate
FROM ONB2ProcessTask AS t
LEFT JOIN User AS u ON t.process.processTrigger.userId = u.userId
LEFT JOIN EmpJob AS ej ON u.userId = ej.userId AND ej.isPrimary = 'true'
WHERE t.lastModifiedDate BETWEEN '{StartDate}' AND '{EndDate}' AND ActivityName IS NOT NULL

UNION ALL

-- Block 4: From the ONB2Process based Integration
SELECT
  p.processId AS OnboardingCaseId,
  'Onboarding Process Completed' AS ActivityName,
  p.lastModifiedDate AS EventTimestamp, -- Assuming last modification on a completed status marks the end
  'SAP SuccessFactors' AS SourceSystem,
  GETDATE() AS LastDataUpdateTimestamp,
  p.lastModifiedDate AS EndTime,
  u.userId AS NewHireEmployeeId,
  ej.department AS Department,
  ej.position AS PositionTitle,
  ej.managerId AS HiringManager,
  p.processor AS OnboardingCoordinator,
  p.targetDate AS OfficialStartDate
FROM ONB2Process AS p
LEFT JOIN User AS u ON p.processTrigger.userId = u.userId
LEFT JOIN EmpJob AS ej ON u.userId = ej.userId AND ej.isPrimary = 'true'
WHERE p.processStatus = 'COMPLETED' AND p.lastModifiedDate BETWEEN '{StartDate}' AND '{EndDate}';

import requests
import pandas as pd
from datetime import datetime
import json
import sys

class=class="hl-string">"hl-comment"># --- Configuration: Replace with your SuccessFactors details ---
SF_API_CENTER_URL = class="hl-string">"https://apiX.successfactors.com/odata/v2" class=class="hl-string">"hl-comment"># e.g., api2.successfactors.eu
COMPANY_ID = class="hl-string">"[Your Company ID]"
API_USERNAME = class="hl-string">"[Your API Username]"
API_PASSWORD = class="hl-string">"[Your API Password]"

class=class="hl-string">"hl-comment"># --- Date Range for Extraction ---
start_date = class="hl-string">"2023-01-01T00:00:00Z"
end_date = class="hl-string">"2023-12-31T23:59:59Z"

class=class="hl-string">"hl-comment"># --- Customizable Task Names from your Onboarding Process ---
TASK_NAMES = {
    class="hl-string">"BACKGROUND_CHECK_INITIATED": class="hl-string">"[Your Background Check Task Name]",
    class="hl-string">"BACKGROUND_CHECK_COMPLETED": class="hl-string">"[Your Background Check Task Name]",
    class="hl-string">"MANAGER_TASKS_ASSIGNED": class="hl-string">"[Your Manager Welcome Tasks Group Name]",
    class="hl-string">"IT_PROVISIONING_REQUESTED": class="hl-string">"[Your IT Provisioning Task Name]",
    class="hl-string">"EQUIPMENT_ORDERED": class="hl-string">"[Your Equipment Order Task Name]",
    class="hl-string">"EQUIPMENT_DELIVERED": class="hl-string">"[Your Equipment Delivery Task Name]",
    class="hl-string">"TRAINING_ASSIGNED": class="hl-string">"[Your Mandatory Training Task Name]",
    class="hl-string">"TRAINING_COMPLETED": class="hl-string">"[Your Mandatory Training Task Name]",
    class="hl-string">"30_DAY_CHECKIN": class="hl-string">"[Your 30-Day Check-in Task Name]"
}

class=class="hl-string">"hl-comment"># --- Helper function for paginated OData calls ---
def fetch_odata_records(endpoint_url):
    records = []
    url = endpoint_url
    print(fclass="hl-string">"Fetching from: {url.split('?')[0]}...")
    while url:
        try:
            response = requests.get(url, auth=(fclass="hl-string">"{API_USERNAME}@{COMPANY_ID}", API_PASSWORD), headers={class="hl-string">'Accept': class="hl-string">'application/json'})
            response.raise_for_status()  class=class="hl-string">"hl-comment"># Raise HTTPError for bad responses (4xx or 5xx)
            data = response.json()[class="hl-string">'d']
            records.extend(data.get(class="hl-string">'results', [data])) class=class="hl-string">"hl-comment"># Handle single and multiple results
            url = data.get(class="hl-string">'__next', None)
        except requests.exceptions.RequestException as e:
            print(fclass="hl-string">"Error fetching data: {e}", file=sys.stderr)
            if response:
                print(fclass="hl-string">"Response content: {response.text}", file=sys.stderr)
            return [] class=class="hl-string">"hl-comment"># Return empty on error
    print(fclass="hl-string">"Fetched {len(records)} records.")
    return records

class=class="hl-string">"hl-comment"># --- Main Extraction Logic ---
def main():
    all_events = []
    last_update_ts = datetime.utcnow().isoformat() + class="hl-string">'Z'

    class=class="hl-string">"hl-comment"># Base query for ONB2Process to get main case data
    process_query_url = (
        fclass="hl-string">"{SF_API_CENTER_URL}/ONB2Process?" 
        fclass="hl-string">"$select=processId,createdDateTime,processStatus,completionDate,subjectUser,hiringManager,onboardingCoordinator,onboardingExternalUser/startDate,subjectUser/userId,subjectUser/department,subjectUser/jobTitle&" 
        fclass="hl-string">"$expand=subjectUser,hiringManager,onboardingCoordinator,onboardingExternalUser&" 
        fclass="hl-string">"$filter=createdDateTime ge datetime'{start_date}' and createdDateTime le datetime'{end_date}'"
    )
    processes = fetch_odata_records(process_query_url)
    if not processes:
        print(class="hl-string">"No onboarding processes found for the given date range. Exiting.")
        return

    process_df = pd.json_normalize(processes)
    class=class="hl-string">"hl-comment"># Rename columns for clarity
    process_df.rename(columns={
        class="hl-string">'processId': class="hl-string">'OnboardingCaseId',
        class="hl-string">'subjectUser.userId': class="hl-string">'NewHireEmployeeId',
        class="hl-string">'subjectUser.department': class="hl-string">'Department',
        class="hl-string">'subjectUser.jobTitle': class="hl-string">'PositionTitle',
        class="hl-string">'hiringManager.fullName': class="hl-string">'HiringManager',
        class="hl-string">'onboardingCoordinator.fullName': class="hl-string">'OnboardingCoordinator',
        class="hl-string">'onboardingExternalUser.startDate': class="hl-string">'OfficialStartDate'
    }, inplace=True)
    
    class=class="hl-string">"hl-comment"># Ensure OfficialStartDate is a clean date string if it exists
    if class="hl-string">'OfficialStartDate' in process_df.columns:
        process_df[class="hl-string">'OfficialStartDate'] = process_df[class="hl-string">'OfficialStartDate'].apply(lambda x: pd.to_datetime(int(x.replace(class="hl-string">'/Date(',class="hl-string">'').replace(class="hl-string">')/',class="hl-string">'')), unit=class="hl-string">'ms').strftime(class="hl-string">'%Y-%m-%dT%H:%M:%SZ') if pd.notna(x) else None)

    class=class="hl-string">"hl-comment"># 1. Onboarding Process Initiated
    df = process_df[[class="hl-string">'OnboardingCaseId', class="hl-string">'createdDateTime']].copy()
    df[class="hl-string">'ActivityName'] = class="hl-string">'Onboarding Process Initiated'
    df.rename(columns={class="hl-string">'createdDateTime': class="hl-string">'EventTimestamp'}, inplace=True)
    all_events.append(df)

    class=class="hl-string">"hl-comment"># 2. Pre-Day 1 Paperwork Sent (Approximated by process initiation)
    df = process_df[[class="hl-string">'OnboardingCaseId', class="hl-string">'createdDateTime']].copy()
    df[class="hl-string">'ActivityName'] = class="hl-string">'Pre-Day 1 Paperwork Sent'
    df.rename(columns={class="hl-string">'createdDateTime': class="hl-string">'EventTimestamp'}, inplace=True)
    all_events.append(df)

    class=class="hl-string">"hl-comment"># 5. Pre-Day 1 Paperwork Completed
    class=class="hl-string">"hl-comment"># Assuming this corresponds to a specific process status change
    df = process_df.loc[process_df[class="hl-string">'processStatus'] == class="hl-string">'ONB2ProcessStatus/READY_TO_HIRE', [class="hl-string">'OnboardingCaseId', class="hl-string">'completionDate']].copy()
    if not df.empty:
        df[class="hl-string">'ActivityName'] = class="hl-string">'Pre-Day 1 Paperwork Completed'
        df.rename(columns={class="hl-string">'completionDate': class="hl-string">'EventTimestamp'}, inplace=True)
        all_events.append(df)

    class=class="hl-string">"hl-comment"># 6. HR Profile Created (Approximated by completion of class="hl-string">'Manage Pending Hires' step)
    df = process_df.loc[process_df[class="hl-string">'processStatus'] == class="hl-string">'ONB2ProcessStatus/HIRED', [class="hl-string">'OnboardingCaseId', class="hl-string">'completionDate']].copy()
    if not df.empty:
        df[class="hl-string">'ActivityName'] = class="hl-string">'HR Profile Created'
        df.rename(columns={class="hl-string">'completionDate': class="hl-string">'EventTimestamp'}, inplace=True)
        all_events.append(df)

    class=class="hl-string">"hl-comment"># 11. First Day Occurred
    df = process_df[[class="hl-string">'OnboardingCaseId', class="hl-string">'OfficialStartDate']].dropna().copy()
    df[class="hl-string">'ActivityName'] = class="hl-string">'First Day Occurred'
    df.rename(columns={class="hl-string">'OfficialStartDate': class="hl-string">'EventTimestamp'}, inplace=True)
    all_events.append(df)
    
    class=class="hl-string">"hl-comment"># 15. Onboarding Process Completed
    df = process_df.loc[process_df[class="hl-string">'processStatus'] == class="hl-string">'ONB2ProcessStatus/COMPLETED', [class="hl-string">'OnboardingCaseId', class="hl-string">'completionDate']].copy()
    if not df.empty:
        df[class="hl-string">'ActivityName'] = class="hl-string">'Onboarding Process Completed'
        df.rename(columns={class="hl-string">'completionDate': class="hl-string">'EventTimestamp'}, inplace=True)
        all_events.append(df)

    class=class="hl-string">"hl-comment"># Fetch all tasks for the relevant processes
    process_ids_str = class="hl-string">"'" + class="hl-string">"','".join(process_df[class="hl-string">'OnboardingCaseId'].astype(str)) + class="hl-string">"'"
    tasks_query_url = (
        fclass="hl-string">"{SF_API_CENTER_URL}/ONB2ProcessTask?"
        fclass="hl-string">"$select=process/processId,taskName,createdDate,completedDate,taskStatus&"
        fclass="hl-string">"$expand=process&"
        fclass="hl-string">"$filter=process/processId in [{process_ids_str}]"
    )
    tasks = fetch_odata_records(tasks_query_url)
    
    if tasks:
        tasks_df = pd.json_normalize(tasks)
        tasks_df.rename(columns={class="hl-string">'process.processId': class="hl-string">'OnboardingCaseId'}, inplace=True)

        def extract_task_events(task_name_key, activity_name, timestamp_col):
            task_name = TASK_NAMES[task_name_key]
            filter_condition = (tasks_df[class="hl-string">'taskName'] == task_name)
            if timestamp_col == class="hl-string">'completedDate':
                 filter_condition &= (tasks_df[class="hl-string">'taskStatus'] == class="hl-string">'ONB2TaskStatus/COMPLETED')

            df = tasks_df.loc[filter_condition, [class="hl-string">'OnboardingCaseId', timestamp_col]].copy()
            if not df.empty:
                df[class="hl-string">'ActivityName'] = activity_name
                df.rename(columns={timestamp_col: class="hl-string">'EventTimestamp'}, inplace=True)
                return df
            return None

        class=class="hl-string">"hl-comment"># 3. Background Check Initiated
        event_df = extract_task_events(class="hl-string">'BACKGROUND_CHECK_INITIATED', class="hl-string">'Background Check Initiated', class="hl-string">'createdDate')
        if event_df is not None: all_events.append(event_df)

        class=class="hl-string">"hl-comment"># 4. Background Check Completed
        event_df = extract_task_events(class="hl-string">'BACKGROUND_CHECK_COMPLETED', class="hl-string">'Background Check Completed', class="hl-string">'completedDate')
        if event_df is not None: all_events.append(event_df)

        class=class="hl-string">"hl-comment"># 7. Manager Tasks Assigned
        event_df = extract_task_events(class="hl-string">'MANAGER_TASKS_ASSIGNED', class="hl-string">'Manager Tasks Assigned', class="hl-string">'createdDate')
        if event_df is not None: all_events.append(event_df)

        class=class="hl-string">"hl-comment"># 8. IT Provisioning Requested
        event_df = extract_task_events(class="hl-string">'IT_PROVISIONING_REQUESTED', class="hl-string">'IT Provisioning Requested', class="hl-string">'createdDate')
        if event_df is not None: all_events.append(event_df)

        class=class="hl-string">"hl-comment"># 9. Equipment Ordered
        event_df = extract_task_events(class="hl-string">'EQUIPMENT_ORDERED', class="hl-string">'Equipment Ordered', class="hl-string">'createdDate')
        if event_df is not None: all_events.append(event_df)

        class=class="hl-string">"hl-comment"># 10. Equipment Delivered
        event_df = extract_task_events(class="hl-string">'EQUIPMENT_DELIVERED', class="hl-string">'Equipment Delivered', class="hl-string">'completedDate')
        if event_df is not None: all_events.append(event_df)

        class=class="hl-string">"hl-comment"># 12. Mandatory Training Assigned
        event_df = extract_task_events(class="hl-string">'TRAINING_ASSIGNED', class="hl-string">'Mandatory Training Assigned', class="hl-string">'createdDate')
        if event_df is not None: all_events.append(event_df)

        class=class="hl-string">"hl-comment"># 13. Mandatory Training Completed
        event_df = extract_task_events(class="hl-string">'TRAINING_COMPLETED', class="hl-string">'Mandatory Training Completed', class="hl-string">'completedDate')
        if event_df is not None: all_events.append(event_df)

        class=class="hl-string">"hl-comment"># 14. 30-Day Check-in Performed
        event_df = extract_task_events(class="hl-string">'30_DAY_CHECKIN', class="hl-string">'30-Day Check-in Performed', class="hl-string">'completedDate')
        if event_df is not None: all_events.append(event_df)

    class=class="hl-string">"hl-comment"># Combine all events
    if not all_events:
        print(class="hl-string">"No events were generated. Please check your configuration and API responses.")
        return

    final_df = pd.concat(all_events, ignore_index=True)
    final_df.dropna(subset=[class="hl-string">'EventTimestamp', class="hl-string">'OnboardingCaseId'], inplace=True)
    
    class=class="hl-string">"hl-comment"># Clean timestamp format
    final_df[class="hl-string">'EventTimestamp'] = final_df[class="hl-string">'EventTimestamp'].apply(lambda x: pd.to_datetime(int(x.replace(class="hl-string">'/Date(',class="hl-string">'').replace(class="hl-string">')/',class="hl-string">'')), unit=class="hl-string">'ms').strftime(class="hl-string">'%Y-%m-%dT%H:%M:%SZ') if isinstance(x, str) and class="hl-string">'/Date(' in x else x)
    final_df[class="hl-string">'EventTimestamp'] = pd.to_datetime(final_df[class="hl-string">'EventTimestamp']).dt.strftime(class="hl-string">'%Y-%m-%dT%H:%M:%S.%f')[:-3] + class="hl-string">'Z'

    class=class="hl-string">"hl-comment"># Merge with process attributes
    attributes_to_merge = [
        class="hl-string">'OnboardingCaseId', class="hl-string">'NewHireEmployeeId', class="hl-string">'Department', class="hl-string">'PositionTitle',
        class="hl-string">'HiringManager', class="hl-string">'OnboardingCoordinator', class="hl-string">'OfficialStartDate'
    ]
    final_df = pd.merge(final_df, process_df[attributes_to_merge], on=class="hl-string">'OnboardingCaseId', how=class="hl-string">'left')
    final_df[class="hl-string">'EndTime'] = final_df[class="hl-string">'EventTimestamp']
    final_df[class="hl-string">'SourceSystem'] = class="hl-string">'SAP SuccessFactors Onboarding'
    final_df[class="hl-string">'LastDataUpdateTimestamp'] = last_update_ts

    class=class="hl-string">"hl-comment"># Reorder columns
    final_df = final_df[[
        class="hl-string">'OnboardingCaseId', class="hl-string">'ActivityName', class="hl-string">'EventTimestamp', class="hl-string">'EndTime',
        class="hl-string">'NewHireEmployeeId', class="hl-string">'Department', class="hl-string">'PositionTitle', class="hl-string">'HiringManager',
        class="hl-string">'OnboardingCoordinator', class="hl-string">'OfficialStartDate', class="hl-string">'SourceSystem', class="hl-string">'LastDataUpdateTimestamp'
    ]]

    class=class="hl-string">"hl-comment"># Save to CSV
    final_df.to_csv(class="hl-string">'sfsf_onboarding_event_log.csv', index=False)
    print(class="hl-string">"\nExtraction complete. Event log saved to 'sfsf_onboarding_event_log.csv'")

if __name__ == class="hl-string">'__main__':
    main()