Mẫu dữ liệu xử lý thanh toán khoản phải trả

/* 
  POST-EXTRACTION TRANSFORMATION SCRIPT
  Apply this SQL logic to the raw Excel export from SQVI to generate the Event Log.
  Input Table Name: [Raw_SQVI_Export]
*/

/* 1. Invoice Receipt Recorded */
SELECT
  "Invoice Number" AS InvoiceNumber,
  'Invoice Receipt Recorded' AS ActivityName,
  CAST("Document Date" AS DATETIME) AS EventTimestamp,
  'SAP ECC' AS SourceSystem,
  GETDATE() AS LastDataUpdate,
  "Company Code" AS CompanyCode,
  "Vendor" AS VendorNumber,
  "Amount" AS InvoiceAmount,
  "Currency" AS Currency
FROM [Raw_SQVI_Export]
WHERE "Document Date" IS NOT NULL

UNION ALL

/* 2. Invoice Created */
SELECT
  "Invoice Number",
  'Invoice Created',
  CAST(CONCAT("Entry Date", ' ', "Entry Time") AS DATETIME),
  'SAP ECC',
  GETDATE(),
  "Company Code",
  "Vendor",
  "Amount",
  "Currency"
FROM [Raw_SQVI_Export]
WHERE "Entry Date" IS NOT NULL

UNION ALL

/* 3. Invoice Parked */
/* Limitation: Historical park status is not preserved in BSAK. Utilizing Entry Date as proxy if distinct from Posting. */
SELECT
  "Invoice Number",
  'Invoice Parked',
  CAST(CONCAT("Entry Date", ' ', "Entry Time") AS DATETIME),
  'SAP ECC',
  GETDATE(),
  "Company Code",
  "Vendor",
  "Amount",
  "Currency"
FROM [Raw_SQVI_Export]
WHERE "Entry Date" < "Posting Date"

UNION ALL

/* 4. Invoice Posted */
SELECT
  "Invoice Number",
  'Invoice Posted',
  CAST("Posting Date" AS DATETIME),
  'SAP ECC',
  GETDATE(),
  "Company Code",
  "Vendor",
  "Amount",
  "Currency"
FROM [Raw_SQVI_Export]
WHERE "Posting Date" IS NOT NULL

UNION ALL

/* 5. Payment Block Applied */
/* Limitation: Only shows current block status (ZLSPR) on the cleared item. */
SELECT
  "Invoice Number",
  'Payment Block Applied',
  CAST("Posting Date" AS DATETIME),
  'SAP ECC',
  GETDATE(),
  "Company Code",
  "Vendor",
  "Amount",
  "Currency"
FROM [Raw_SQVI_Export]
WHERE "Payment Block" IS NOT NULL AND "Payment Block" <> ''

UNION ALL

/* 6. Invoice Data Changed */
/* Limitation: Cannot be extracted via SQVI BKPF/BSAK join. Requires CDHDR table. Row excluded or placeholder. */
SELECT
  "Invoice Number",
  'Invoice Data Changed',
  NULL,
  'SAP ECC',
  GETDATE(),
  "Company Code",
  "Vendor",
  "Amount",
  "Currency"
FROM [Raw_SQVI_Export]
WHERE 1=0

UNION ALL

/* 7. Payment Term Changed */
/* Limitation: Requires Change Log extraction. */
SELECT
  "Invoice Number",
  'Payment Term Changed',
  NULL,
  'SAP ECC',
  GETDATE(),
  "Company Code",
  "Vendor",
  "Amount",
  "Currency"
FROM [Raw_SQVI_Export]
WHERE 1=0

UNION ALL

/* 8. Payment Block Removed */
/* Limitation: Requires Change Log extraction. */
SELECT
  "Invoice Number",
  'Payment Block Removed',
  NULL,
  'SAP ECC',
  GETDATE(),
  "Company Code",
  "Vendor",
  "Amount",
  "Currency"
FROM [Raw_SQVI_Export]
WHERE 1=0

UNION ALL

/* 9. Payment Proposal Created */
/* Proxy: Uses Clearing Date minus 1 day as estimation since REGUH table is not in join. */
SELECT
  "Invoice Number",
  'Payment Proposal Created',
  DATEADD(day, -1, CAST("Clearing Date" AS DATETIME)),
  'SAP ECC',
  GETDATE(),
  "Company Code",
  "Vendor",
  "Amount",
  "Currency"
FROM [Raw_SQVI_Export]
WHERE "Clearing Date" IS NOT NULL

UNION ALL

/* 10. Payment Posted */
/* Logic: Uses Clearing Date as the effective payment date. */
SELECT
  "Invoice Number",
  'Payment Posted',
  CAST("Clearing Date" AS DATETIME),
  'SAP ECC',
  GETDATE(),
  "Company Code",
  "Vendor",
  "Amount",
  "Currency"
FROM [Raw_SQVI_Export]
WHERE "Clearing Date" IS NOT NULL AND "Clearing Doc" IS NOT NULL

UNION ALL

/* 11. Invoice Cleared */
SELECT
  "Invoice Number",
  'Invoice Cleared',
  CAST("Clearing Date" AS DATETIME),
  'SAP ECC',
  GETDATE(),
  "Company Code",
  "Vendor",
  "Amount",
  "Currency"
FROM [Raw_SQVI_Export]
WHERE "Clearing Date" IS NOT NULL

UNION ALL

/* 12. Discount Due Date Passed */
/* Logic: Checks if Clearing Date > (Baseline Date + Term 1 Days). */
SELECT
  "Invoice Number",
  'Discount Due Date Passed',
  DATEADD(day, CAST("Terms 1" AS INT), CAST("Baseline Date" AS DATETIME)),
  'SAP ECC',
  GETDATE(),
  "Company Code",
  "Vendor",
  "Amount",
  "Currency"
FROM [Raw_SQVI_Export]
WHERE "Clearing Date" > DATEADD(day, CAST("Terms 1" AS INT), CAST("Baseline Date" AS DATETIME))

UNION ALL

/* 13. Due Date Passed */
/* Logic: Checks if Clearing Date > Net Due Date. */
SELECT
  "Invoice Number",
  'Due Date Passed',
  DATEADD(day, CAST("Terms 1" AS INT) + CAST("Terms 2" AS INT), CAST("Baseline Date" AS DATETIME)),
  'SAP ECC',
  GETDATE(),
  "Company Code",
  "Vendor",
  "Amount",
  "Currency"
FROM [Raw_SQVI_Export]
WHERE "Clearing Date" > DATEADD(day, CAST("Terms 1" AS INT) + CAST("Terms 2" AS INT), CAST("Baseline Date" AS DATETIME))

import sys
from datetime import datetime, timedelta
import pandas as pd
from pyrfc import Connection

class=class="hl-string">"hl-comment"># === CONFIGURATION ===
SAP_CONN_PARAMS = {
    class="hl-string">'user': class="hl-string">'YOUR_USER',
    class="hl-string">'passwd': class="hl-string">'YOUR_PASSWORD',
    class="hl-string">'ashost': class="hl-string">'YOUR_SAP_HOST',
    class="hl-string">'sysnr': class="hl-string">'00',
    class="hl-string">'client': class="hl-string">'100',
    class="hl-string">'lang': class="hl-string">'EN'
}

TARGET_COMPANY_CODE = class="hl-string">'1000'
DATE_FROM = class="hl-string">'20230101'
DATE_TO = class="hl-string">'20230131'
SOURCE_SYSTEM_ID = class="hl-string">'SAP_ECC_01'

class=class="hl-string">"hl-comment"># === RFC HELPER ===
def read_table(conn, table_name, fields, options, row_limit=0):
    class="hl-string">"""Reads SAP table via RFC_READ_TABLE and returns a DataFrame."""
    print(fclass="hl-string">"Reading {table_name}...")
    
    class=class="hl-string">"hl-comment"># Format fields for RFC
    rfc_fields = [{class="hl-string">'FIELDNAME': f} for f in fields]
    
    class=class="hl-string">"hl-comment"># Call RFC
    try:
        result = conn.call(class="hl-string">'RFC_READ_TABLE',
                           QUERY_TABLE=table_name,
                           DELIMITER=class="hl-string">';',
                           FIELDS=rfc_fields,
                           OPTIONS=[{class="hl-string">'TEXT': o} for o in options],
                           ROWCOUNT=row_limit)
    except Exception as e:
        print(fclass="hl-string">"Error reading {table_name}: {e}")
        return pd.DataFrame()

    class=class="hl-string">"hl-comment"># Parse output
    data = []
    headers = [f[class="hl-string">'FIELDNAME'] for f in result[class="hl-string">'FIELDS']]
    
    for line in result[class="hl-string">'DATA']:
        class=class="hl-string">"hl-comment"># Simple split by delimiter; robust parsing handles fixed width if delimiter fails, 
        class=class="hl-string">"hl-comment"># but semicolon is standard for this approach.
        values = line[class="hl-string">'WA'].split(class="hl-string">';')
        class=class="hl-string">"hl-comment"># Handle cases where split might be shorter if trailing fields are empty
        if len(values) < len(headers):
            values += [class="hl-string">''] * (len(headers) - len(values))
        data.append(dict(zip(headers, values)))
        
    return pd.DataFrame(data)

def main():
    class=class="hl-string">"hl-comment"># 1. Connect to SAP
    try:
        conn = Connection(**SAP_CONN_PARAMS)
    except Exception as e:
        print(fclass="hl-string">"Connection failed: {e}")
        return

    class=class="hl-string">"hl-comment"># 2. Extract BKPF (Header)
    class=class="hl-string">"hl-comment"># Filters: Company Code and Entry Date (CPUDT)
    bkpf_fields = [class="hl-string">'BUKRS', class="hl-string">'BELNR', class="hl-string">'GJAHR', class="hl-string">'BLART', class="hl-string">'BLDAT', class="hl-string">'BUDAT', class="hl-string">'CPUDT', class="hl-string">'CPUTM', class="hl-string">'USNAM']
    bkpf_opts = [
        fclass="hl-string">"BUKRS = '{TARGET_COMPANY_CODE}' AND CPUDT >= '{DATE_FROM}' AND CPUDT <= '{DATE_TO}'"
    ]
    df_bkpf = read_table(conn, class="hl-string">'BKPF', bkpf_fields, bkpf_opts)
    
    if df_bkpf.empty:
        print(class="hl-string">"No documents found.")
        return

    class=class="hl-string">"hl-comment"># 3. Extract Vendor Line Items (BSAK - Cleared, BSIK - Open)
    class=class="hl-string">"hl-comment"># To avoid huge queries, we filter BSIK/BSAK by the same date range or join later.
    class=class="hl-string">"hl-comment"># For robustness in RFC, we filter by BUKRS and Date range on CPU date or BUDAT similar to BKPF
    item_fields = [class="hl-string">'BUKRS', class="hl-string">'BELNR', class="hl-string">'GJAHR', class="hl-string">'BUZEI', class="hl-string">'LIFNR', class="hl-string">'DMBTR', class="hl-string">'WAERS', class="hl-string">'ZFBDT', class="hl-string">'ZTERM', class="hl-string">'ZBD1T', class="hl-string">'ZBD2T', class="hl-string">'ZLSPR', class="hl-string">'AUGDT', class="hl-string">'AUGBL']
    
    class=class="hl-string">"hl-comment"># Note: BSIK/BSAK are huge. In production, chunk by BELNR list derived from BKPF or Date.
    class=class="hl-string">"hl-comment"># Here we use the date range on BUDAT for simplicity of the example.
    item_opts = [
        fclass="hl-string">"BUKRS = '{TARGET_COMPANY_CODE}' AND BUDAT >= '{DATE_FROM}' AND BUDAT <= '{DATE_TO}'"
    ]
    
    df_bsik = read_table(conn, class="hl-string">'BSIK', item_fields, item_opts)
    df_bsak = read_table(conn, class="hl-string">'BSAK', item_fields, item_opts)
    
    class=class="hl-string">"hl-comment"># Combine Open and Cleared items
    df_items = pd.concat([df_bsik, df_bsak], ignore_index=True)
    
    class=class="hl-string">"hl-comment"># Filter items to only those present in our BKPF fetch (Inner Join logic)
    df_items = df_items[df_items[class="hl-string">'BELNR'].isin(df_bkpf[class="hl-string">'BELNR'])]

    class=class="hl-string">"hl-comment"># 4. Extract Payment Data (REGUH)
    class=class="hl-string">"hl-comment"># Filter by Date range roughly matching the process
    reguh_fields = [class="hl-string">'LAUFD', class="hl-string">'LAUFI', class="hl-string">'XVORL', class="hl-string">'VBLNR', class="hl-string">'LIFNR']
    reguh_opts = [
        fclass="hl-string">"LAUFD >= '{DATE_FROM}'"
    ]
    df_reguh = read_table(conn, class="hl-string">'REGUH', reguh_fields, reguh_opts)

    class=class="hl-string">"hl-comment"># 5. Extract Change History (CDHDR/CDPOS)
    class=class="hl-string">"hl-comment"># This is expensive. Filter by OBJECTCLAS class="hl-string">'BELEG' and Date.
    cdhdr_fields = [class="hl-string">'OBJECTCLAS', class="hl-string">'OBJECTID', class="hl-string">'CHANGENR', class="hl-string">'UDATE', class="hl-string">'UTIME', class="hl-string">'USERNAME']
    cdhdr_opts = [
        fclass="hl-string">"OBJECTCLAS = 'BELEG' AND UDATE >= '{DATE_FROM}'"
    ]
    df_cdhdr = read_table(conn, class="hl-string">'CDHDR', cdhdr_fields, cdhdr_opts)
    
    class=class="hl-string">"hl-comment"># Filter CDHDR to only relevant Invoice Numbers (OBJECTID usually contains BELNR + GJAHR)
    class=class="hl-string">"hl-comment"># Construct OBJECTID list from BKPF
    valid_object_ids = set(df_bkpf[class="hl-string">'BELNR'] + df_bkpf[class="hl-string">'GJAHR'])
    df_cdhdr = df_cdhdr[df_cdhdr[class="hl-string">'OBJECTID'].isin(valid_object_ids)]
    
    if not df_cdhdr.empty:
        class=class="hl-string">"hl-comment"># Get CDPOS for these changes
        class=class="hl-string">"hl-comment"># Cannot pass thousands of CHANGENR. Split if needed. 
        class=class="hl-string">"hl-comment"># Here assuming volume allows single fetch or loop.
        change_nums = df_cdhdr[class="hl-string">'CHANGENR'].unique().tolist()
        class=class="hl-string">"hl-comment"># Chunking needed for production
        cdpos_fields = [class="hl-string">'OBJECTCLAS', class="hl-string">'OBJECTID', class="hl-string">'CHANGENR', class="hl-string">'TABNAME', class="hl-string">'FNAME', class="hl-string">'VALUE_NEW', class="hl-string">'VALUE_OLD']
        df_cdpos = pd.DataFrame()
        
        class=class="hl-string">"hl-comment"># Simple chunk loop
        chunk_size = 100
        for i in range(0, len(change_nums), chunk_size):
            chunk = change_nums[i:i + chunk_size]
            chunk_str = class="hl-string">"'".join(chunk)
            opts = [fclass="hl-string">"OBJECTCLAS = 'BELEG' AND CHANGENR IN ('{chunk_str}')"]
            df_chunk = read_table(conn, class="hl-string">'CDPOS', cdpos_fields, opts)
            df_cdpos = pd.concat([df_cdpos, df_chunk])
            
        class=class="hl-string">"hl-comment"># Merge CDHDR and CDPOS
        df_changes = pd.merge(df_cdpos, df_cdhdr, on=[class="hl-string">'OBJECTCLAS', class="hl-string">'OBJECTID', class="hl-string">'CHANGENR'])
    else:
        df_changes = pd.DataFrame(columns=[class="hl-string">'OBJECTID', class="hl-string">'FNAME', class="hl-string">'VALUE_NEW', class="hl-string">'VALUE_OLD', class="hl-string">'UDATE', class="hl-string">'UTIME', class="hl-string">'USERNAME'])

    class=class="hl-string">"hl-comment"># === PROCESS EVENTS ===
    events = []

    class=class="hl-string">"hl-comment"># Join Headers and Items
    class=class="hl-string">"hl-comment"># Ensure key columns are strings for merging
    df_full = pd.merge(df_bkpf, df_items, on=[class="hl-string">'BUKRS', class="hl-string">'BELNR', class="hl-string">'GJAHR'])

    for _, row in df_full.iterrows():
        invoice_num = row[class="hl-string">'BELNR']
        vendor = row[class="hl-string">'LIFNR']
        amount = row[class="hl-string">'DMBTR']
        
        class=class="hl-string">"hl-comment"># Common attributes
        base_attrs = {
            class="hl-string">'InvoiceNumber': invoice_num,
            class="hl-string">'SourceSystem': SOURCE_SYSTEM_ID,
            class="hl-string">'CompanyCode': row[class="hl-string">'BUKRS'],
            class="hl-string">'VendorNumber': vendor,
            class="hl-string">'InvoiceAmount': amount,
            class="hl-string">'DocumentType': row[class="hl-string">'BLART'],
            class="hl-string">'LastDataUpdate': datetime.now().isoformat()
        }

        class=class="hl-string">"hl-comment"># 1. Invoice Created
        events.append({
            **base_attrs,
            class="hl-string">'ActivityName': class="hl-string">'Invoice Created',
            class="hl-string">'EventTimestamp': fclass="hl-string">"{row['CPUDT']}T{row['CPUTM']}",
            class="hl-string">'UserName': row[class="hl-string">'USNAM']
        })

        class=class="hl-string">"hl-comment"># 2. Invoice Receipt Recorded
        if row[class="hl-string">'BLDAT']:
            events.append({
                **base_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Invoice Receipt Recorded',
                class="hl-string">'EventTimestamp': fclass="hl-string">"{row['BLDAT']}T00:00:00",
                class="hl-string">'UserName': row[class="hl-string">'USNAM']
            })

        class=class="hl-string">"hl-comment"># 3. Invoice Posted
        if row[class="hl-string">'BUDAT']:
            events.append({
                **base_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Invoice Posted',
                class="hl-string">'EventTimestamp': fclass="hl-string">"{row['BUDAT']}T00:00:00",
                class="hl-string">'UserName': row[class="hl-string">'USNAM']
            })

        class=class="hl-string">"hl-comment"># 4. Invoice Cleared
        if row[class="hl-string">'AUGDT'] and row[class="hl-string">'AUGDT'] != class="hl-string">'00000000':
            events.append({
                **base_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Invoice Cleared',
                class="hl-string">'EventTimestamp': fclass="hl-string">"{row['AUGDT']}T00:00:00",
                class="hl-string">'ClearingDate': row[class="hl-string">'AUGDT']
            })
            
            class=class="hl-string">"hl-comment"># 5. Payment Posted (Proxy via Clearing)
            class=class="hl-string">"hl-comment"># If cleared by payment document, assume Payment Posted same day
            events.append({
                **base_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Payment Posted',
                class="hl-string">'EventTimestamp': fclass="hl-string">"{row['AUGDT']}T23:59:59"
            })

        class=class="hl-string">"hl-comment"># 6. Payment Proposal Created
        class=class="hl-string">"hl-comment"># Check REGUH
        class=class="hl-string">"hl-comment"># Logic: REGUH contains payment proposals. We link via VBLNR (Payment Doc) or LIFNR matches
        class=class="hl-string">"hl-comment"># Simplified: if we find a proposal for this vendor in the range
        class=class="hl-string">"hl-comment"># In practice, need tighter link via BSAK-AUGBL = REGUH-VBLNR
        if row[class="hl-string">'AUGBL']:
            prop = df_reguh[df_reguh[class="hl-string">'VBLNR'] == row[class="hl-string">'AUGBL']]
            if not prop.empty:
                 for _, p_row in prop.iterrows():
                    events.append({
                        **base_attrs,
                        class="hl-string">'ActivityName': class="hl-string">'Payment Proposal Created',
                        class="hl-string">'EventTimestamp': fclass="hl-string">"{p_row['LAUFD']}T00:00:00"
                    })

        class=class="hl-string">"hl-comment"># 7. Parked (Check if originally parked - simplified logic)
        class=class="hl-string">"hl-comment"># Requires VBKPF check or document status history. Omitted for brevity in this block, 
        class=class="hl-string">"hl-comment"># usually involves checking if document ID exists in VBKPF list.

        class=class="hl-string">"hl-comment"># 8-11. Change Log Activities
        class=class="hl-string">"hl-comment"># Filter changes for this document
        obj_id = row[class="hl-string">'BELNR'] + row[class="hl-string">'GJAHR']
        doc_changes = df_changes[df_changes[class="hl-string">'OBJECTID'] == obj_id]
        
        for _, chg in doc_changes.iterrows():
            ts = fclass="hl-string">"{chg['UDATE']}T{chg['UTIME']}"
            fname = chg[class="hl-string">'FNAME']
            
            if fname == class="hl-string">'ZLSPR':
                if chg[class="hl-string">'VALUE_OLD'] == class="hl-string">'' and chg[class="hl-string">'VALUE_NEW'] != class="hl-string">'':
                    events.append({**base_attrs, class="hl-string">'ActivityName': class="hl-string">'Payment Block Applied', class="hl-string">'EventTimestamp': ts, class="hl-string">'PaymentBlockCode': chg[class="hl-string">'VALUE_NEW']})
                elif chg[class="hl-string">'VALUE_OLD'] != class="hl-string">'' and chg[class="hl-string">'VALUE_NEW'] == class="hl-string">'':
                    events.append({**base_attrs, class="hl-string">'ActivityName': class="hl-string">'Payment Block Removed', class="hl-string">'EventTimestamp': ts})
            elif fname == class="hl-string">'ZTERM':
                events.append({**base_attrs, class="hl-string">'ActivityName': class="hl-string">'Payment Term Changed', class="hl-string">'EventTimestamp': ts})
            else:
                events.append({**base_attrs, class="hl-string">'ActivityName': class="hl-string">'Invoice Data Changed', class="hl-string">'EventTimestamp': ts})

        class=class="hl-string">"hl-comment"># 12. Due Date Passed (Calculated)
        class=class="hl-string">"hl-comment"># Calculate Net Due Date = Baseline Date (ZFBDT) + Max Days (ZBD1T or ZBD2T or ZBD3T)
        try:
            zfbdt = datetime.strptime(row[class="hl-string">'ZFBDT'], class="hl-string">'%Y%m%d')
            days = float(row[class="hl-string">'ZBD1T']) if row[class="hl-string">'ZBD1T'] else 0
            due_date = zfbdt + timedelta(days=days)
            
            class=class="hl-string">"hl-comment"># If today > due_date and not cleared (or cleared after due date)
            clearing_date = datetime.strptime(row[class="hl-string">'AUGDT'], class="hl-string">'%Y%m%d') if (row[class="hl-string">'AUGDT'] and row[class="hl-string">'AUGDT'] != class="hl-string">'00000000') else None
            check_date = clearing_date if clearing_date else datetime.now()
            
            if check_date > due_date:
                 events.append({
                    **base_attrs,
                    class="hl-string">'ActivityName': class="hl-string">'Due Date Passed',
                    class="hl-string">'EventTimestamp': due_date.isoformat()
                })
        except ValueError:
            pass class=class="hl-string">"hl-comment"># Handle date parsing errors

        class=class="hl-string">"hl-comment"># 13. Discount Due Date Passed
        class=class="hl-string">"hl-comment"># Logic similar to Due Date using discount terms

    class=class="hl-string">"hl-comment"># Finalize
    df_events = pd.DataFrame(events)
    print(fclass="hl-string">"Generated {len(df_events)} events.")
    df_events.to_csv(class="hl-string">'sap_ap_event_log.csv', index=False)

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

WITH source_documents AS (
    SELECT
        [Invoice Number] AS InvoiceNumber,
        [Company Code] AS CompanyCode,
        [Vendor Number] AS VendorNumber,
        [Document Type] AS DocumentType,
        [Invoice Amount] AS InvoiceAmount,
        [Purchase Order Number] AS PurchaseOrderNumber,
        [Payment Terms] AS PaymentTerms,
        [Payment Block Code] AS PaymentBlockCode,
        [Payment Method] AS PaymentMethod,
        [User Name] AS UserName,
        [Clearing Date] AS ClearingDate,
        [Document Date] AS DocumentDate,
        [Invoice Created Timestamp] AS InvoiceCreatedTimestamp,
        [Invoice Parked Timestamp] AS InvoiceParkedTimestamp,
        [Invoice Posted Timestamp] AS InvoicePostedTimestamp,
        [Payment Block Applied Timestamp] AS PaymentBlockAppliedTimestamp,
        [Invoice Data Changed Timestamp] AS InvoiceDataChangedTimestamp,
        [Payment Term Changed Timestamp] AS PaymentTermChangedTimestamp,
        [Payment Block Removed Timestamp] AS PaymentBlockRemovedTimestamp,
        [Payment Proposal Created Timestamp] AS PaymentProposalCreatedTimestamp,
        [Payment Posted Timestamp] AS PaymentPostedTimestamp,
        [Invoice Cleared Timestamp] AS InvoiceClearedTimestamp,
        [Discount Due Date 1] AS DiscountDueDate1,
        [Discount Due Date 2] AS DiscountDueDate2,
        [Net Due Date] AS NetDueDate,
        [Source Change Timestamp] AS SourceChangeTimestamp,
        [Source System] AS SourceSystem
    FROM [Your table name]
    WHERE [Source Change Timestamp] >= CAST('[Last successful extraction timestamp]' AS TIMESTAMP)
      AND [Source Change Timestamp] < CAST('[End timestamp]' AS TIMESTAMP)
      AND [Document Date] >= CAST('[Start date]' AS DATE)
      AND [Document Date] < CAST('[End date]' AS DATE)
      AND ([Company Code] IN ([Company Code filter]) OR [Company Code filter] IS NULL)
      AND ([Document Type] IN ([Document Type filter]) OR [Document Type filter] IS NULL)
),
base_events AS (
    SELECT InvoiceNumber, 'Invoice Receipt Recorded' AS ActivityName, CAST(DocumentDate AS TIMESTAMP) AS EventTimestamp, SourceSystem, CompanyCode, VendorNumber, DocumentType, InvoiceAmount, PurchaseOrderNumber, PaymentTerms, PaymentBlockCode, PaymentMethod, UserName, ClearingDate FROM source_documents WHERE DocumentDate IS NOT NULL
    UNION ALL
    SELECT InvoiceNumber, 'Invoice Created', InvoiceCreatedTimestamp, SourceSystem, CompanyCode, VendorNumber, DocumentType, InvoiceAmount, PurchaseOrderNumber, PaymentTerms, PaymentBlockCode, PaymentMethod, UserName, ClearingDate FROM source_documents WHERE InvoiceCreatedTimestamp IS NOT NULL
    UNION ALL
    SELECT InvoiceNumber, 'Invoice Parked', InvoiceParkedTimestamp, SourceSystem, CompanyCode, VendorNumber, DocumentType, InvoiceAmount, PurchaseOrderNumber, PaymentTerms, PaymentBlockCode, PaymentMethod, UserName, ClearingDate FROM source_documents WHERE InvoiceParkedTimestamp IS NOT NULL
    UNION ALL
    SELECT InvoiceNumber, 'Invoice Posted', InvoicePostedTimestamp, SourceSystem, CompanyCode, VendorNumber, DocumentType, InvoiceAmount, PurchaseOrderNumber, PaymentTerms, PaymentBlockCode, PaymentMethod, UserName, ClearingDate FROM source_documents WHERE InvoicePostedTimestamp IS NOT NULL
    UNION ALL
    SELECT InvoiceNumber, 'Payment Block Applied', PaymentBlockAppliedTimestamp, SourceSystem, CompanyCode, VendorNumber, DocumentType, InvoiceAmount, PurchaseOrderNumber, PaymentTerms, PaymentBlockCode, PaymentMethod, UserName, ClearingDate FROM source_documents WHERE PaymentBlockAppliedTimestamp IS NOT NULL
    UNION ALL
    SELECT InvoiceNumber, 'Invoice Data Changed', InvoiceDataChangedTimestamp, SourceSystem, CompanyCode, VendorNumber, DocumentType, InvoiceAmount, PurchaseOrderNumber, PaymentTerms, PaymentBlockCode, PaymentMethod, UserName, ClearingDate FROM source_documents WHERE InvoiceDataChangedTimestamp IS NOT NULL
    UNION ALL
    SELECT InvoiceNumber, 'Payment Term Changed', PaymentTermChangedTimestamp, SourceSystem, CompanyCode, VendorNumber, DocumentType, InvoiceAmount, PurchaseOrderNumber, PaymentTerms, PaymentBlockCode, PaymentMethod, UserName, ClearingDate FROM source_documents WHERE PaymentTermChangedTimestamp IS NOT NULL
    UNION ALL
    SELECT InvoiceNumber, 'Payment Block Removed', PaymentBlockRemovedTimestamp, SourceSystem, CompanyCode, VendorNumber, DocumentType, InvoiceAmount, PurchaseOrderNumber, PaymentTerms, PaymentBlockCode, PaymentMethod, UserName, ClearingDate FROM source_documents WHERE PaymentBlockRemovedTimestamp IS NOT NULL
    UNION ALL
    SELECT InvoiceNumber, 'Payment Proposal Created', PaymentProposalCreatedTimestamp, SourceSystem, CompanyCode, VendorNumber, DocumentType, InvoiceAmount, PurchaseOrderNumber, PaymentTerms, PaymentBlockCode, PaymentMethod, UserName, ClearingDate FROM source_documents WHERE PaymentProposalCreatedTimestamp IS NOT NULL
    UNION ALL
    SELECT InvoiceNumber, 'Payment Posted', PaymentPostedTimestamp, SourceSystem, CompanyCode, VendorNumber, DocumentType, InvoiceAmount, PurchaseOrderNumber, PaymentTerms, PaymentBlockCode, PaymentMethod, UserName, ClearingDate FROM source_documents WHERE PaymentPostedTimestamp IS NOT NULL
    UNION ALL
    SELECT InvoiceNumber, 'Invoice Cleared', InvoiceClearedTimestamp, SourceSystem, CompanyCode, VendorNumber, DocumentType, InvoiceAmount, PurchaseOrderNumber, PaymentTerms, PaymentBlockCode, PaymentMethod, UserName, ClearingDate FROM source_documents WHERE InvoiceClearedTimestamp IS NOT NULL
    UNION ALL
    SELECT InvoiceNumber, 'Discount Due Date Passed', CAST(CURRENT_DATE AS TIMESTAMP), SourceSystem, CompanyCode, VendorNumber, DocumentType, InvoiceAmount, PurchaseOrderNumber, PaymentTerms, PaymentBlockCode, PaymentMethod, UserName, ClearingDate FROM source_documents WHERE ClearingDate IS NULL AND ((DiscountDueDate1 IS NOT NULL AND CURRENT_DATE > DiscountDueDate1) OR (DiscountDueDate2 IS NOT NULL AND CURRENT_DATE > DiscountDueDate2))
    UNION ALL
    SELECT InvoiceNumber, 'Due Date Passed', CAST(CURRENT_DATE AS TIMESTAMP), SourceSystem, CompanyCode, VendorNumber, DocumentType, InvoiceAmount, PurchaseOrderNumber, PaymentTerms, PaymentBlockCode, PaymentMethod, UserName, ClearingDate FROM source_documents WHERE ClearingDate IS NULL AND NetDueDate IS NOT NULL AND CURRENT_DATE > NetDueDate
),
final_events AS (
    SELECT
        InvoiceNumber,
        ActivityName,
        EventTimestamp,
        SourceSystem,
        CAST(CURRENT_TIMESTAMP AS TIMESTAMP) AS LastDataUpdate,
        CompanyCode,
        VendorNumber,
        DocumentType,
        InvoiceAmount,
        PurchaseOrderNumber,
        PaymentTerms,
        PaymentBlockCode,
        PaymentMethod,
        UserName,
        ClearingDate,
        ROW_NUMBER() OVER (PARTITION BY InvoiceNumber, ActivityName, EventTimestamp ORDER BY EventTimestamp) AS DuplicateRank
    FROM base_events
    WHERE InvoiceNumber IS NOT NULL
      AND EventTimestamp IS NOT NULL
)
SELECT
    InvoiceNumber,
    ActivityName,
    EventTimestamp,
    SourceSystem,
    LastDataUpdate,
    CompanyCode,
    VendorNumber,
    DocumentType,
    InvoiceAmount,
    PurchaseOrderNumber,
    PaymentTerms,
    PaymentBlockCode,
    PaymentMethod,
    UserName,
    ClearingDate
FROM final_events
WHERE DuplicateRank = 1
ORDER BY InvoiceNumber, EventTimestamp, ActivityName;