[DATA] version 5.17.1 AZN issues with all balance, income, and cash flow for AZN view="detailed"

[miruddfan]

Data Quality Issue Details

Issue Type:

• [x ] Incorrect financial values (wrong numbers)
• Missing financial data (expected data not present)
• Calculation errors (formulas producing wrong results)
• Data inconsistency (different values for same metric)
• Historical data problems (changes over time)

Environment

EdgarTools Version: 5.17.1
**Python Version:3.10.19 **
Operating System:win 11

Financial Data Details

Company/Ticker: AZN
Form Type: 20-F (e.g., 10-K, 10-Q, 8-K)
**Filing Date/Period:2024 **
Statement Type:Balance Sheet, Income Statement, Cash Flow Statement

Specific Metric/Concept:

• Financial line item: Several
• XBRL concept name: Several

Data Issue

Expected Value:
Provided Excel files with info for all three

Actual Value from EdgarTools:

Code to reproduce:

from edgar import Company

company_ticker = input("Enter company ticker (e.g., MSFT): ").upper()
write_files = input("Do you want to save CSV files of the sheets? Y/N ").upper()
number_years = int(input("Enter the number(years) of annual reports desired = "))

company = Company(company_ticker)
ef = company.get_filings(form=["10-K", "10-K/A"])
filings = ef.latest(number_years)
if filings.latest() is None:
    Taxonomy = 'IFRS'
    filings = company.get_filings(form="20-F", amendments=False).latest(number_years)
    if filings.latest() is None:
       CIK = input("Either the ticker is wrong or it is traded on the OTC.  Please provide the CIK: ")
       company = Company(CIK)
       filings = company.get_filings(form="20-F", amendments=False).latest(number_years)
       if filings is None: 
           print('No SEC Files ')
           sys.exit()
print("Selected filings passed to XBRLS:")
ef_10k = company.get_filings(form=["10-K", "10-K/A"])

if len(ef_10k) == 0:
    print("No 10-K filings found — switching to 20-F (IFRS).")
    Taxonomy = 'IFRS'
    ef_20f = company.get_filings(form="20-F", amendments=False)
    if len(ef_20f) == 0:
        CIK = input("Either the ticker is wrong or it is OTC. Enter the CIK: ")
        company = Company(CIK)
        ef_20f = company.get_filings(form="20-F", amendments=False)
        if len(ef_20f) == 0:
            print("No SEC filings found.")
            sys.exit()
    filings = ef_20f.latest(number_years)

else:
    Taxonomy = 'GAAP'
    filings = ef_10k.latest(number_years)



xbrls = XBRLS.from_filings(filings)

income_statement = xbrls.statements.income_statement(view="detailed")
balance_sheet = xbrls.statements.balance_sheet(view="detailed")
cash_flow = xbrls.statements.cashflow_statement(view="detailed")

bs_df = balance_sheet.to_dataframe()
is_df = income_statement.to_dataframe()
cf_df = cash_flow.to_dataframe()

Cross-Verification

Have you verified this issue with:

• Multiple time periods for same company
• Multiple companies with same issue
• [x ] Direct SEC filing comparison
• Other financial data sources

Affects multiple periods/companies?

• Companies tested: AZN
• Time periods tested: 2024
• Pattern observed: (e.g., all Q4 periods affected, only certain companies)

Expected Behavior

What should happen:

I know you just pushed the recent capability for ifrs, so this is not surprising. I reviewed the output for AZN in 2024 and highlighted information in the excel file. [Unfortunately, I chose a poor color scheme]

yellow = match
red/salmon = error
green = not sure, mainly the number is correct, but the sign is the opposite of what is in the annual report

See the attached excel files

AZN_Balance_Sheets.xlsx
AZN_Income_Statement.xlsx
AZN_Cash_Flow_Statement.xlsx

Data validation rules:

• Should the value be positive/negative?
• Expected magnitude/range?
• Should it match specific calculations?

Additional Context

• (https://www.astrazeneca.com/investor-relations/annual-reports/annual-report-2024.html)
• pages 148 - 151

Impact Assessment:

• Minor (affects specific edge case)
• Moderate (affects common use cases)
• Major (affects core financial calculations)
• Critical (produces completely wrong results)

---

Data quality issues are high priority and will be verified against official SEC filings. Accuracy is fundamental to EdgarTools.

[dgunning]

Thanks for reporting this — we've confirmed the issue.

Root cause: EdgarTools' financial data stitching (MultiFinancials / XBRL stitching) currently has no IFRS concept mappings. AstraZeneca files using IFRS taxonomy (as a foreign private issuer via 20-F), so when view="detailed" tries to normalize and align concepts across periods, it finds no matching mappings and returns empty/missing data.

What's affected:

• get_financials() with view="detailed" for any IFRS filer (AZN, GSK, NVO, SHOP via 40-F, etc.)
• Basic view="standard" may still work since it doesn't require concept normalization

Tracking: This is tracked internally as edgartools-knat, blocked by our broader IFRS stitching improvement effort (edgartools-iq3a). The fix requires:

1. Adding IFRS concept mappings to gaap_mappings.json
2. Creating IFRS ordering templates
3. Updating the concept normalizer to handle IFRS tags

We'll update this issue as progress is made.

[miruddfan]

@dgunning I just ran the same script with view="standard".

Cash Flow: All numbers match the published 10-k, but sign issues need to be cleared up
Income Statement: All numbers match the published 10-k, but sign issues need to be cleared up
Balance Sheet: All numbers match the published 10-k, but sign issues need to be cleared up EXCEPT

•      ifrs-full_CurrentAssets
  

•      ifrs-full_RightofuseAssets
  

•      ifrs-full_NoncurrentDerivativeFinancialAssets
  

•      ifrs-full_NoncurrentAssets
  

•      ifrs-full_Assets
  

I wonder if it has to do with how AZN reports the totals for Non-Current Assets, Current Assets, Non-Current Liabilities, Current Liabilities. The Balance Sheet has no title for these lines...... Not sure how it XBRL reports or handles this information.

[dgunning]

Thanks for testing with view="standard" and providing this detailed breakdown — very helpful.

We've traced the root causes for both issues you're seeing:

Sign errors (all three statements): The IFRS income statement gets resolved internally as ComprehensiveIncome rather than IncomeStatement, which causes the sign-application logic to be skipped entirely. The preferred sign from the XBRL presentation linkbase (which encodes which concepts should display as negated) never gets applied. This also affects how balance type is looked up — the normalization only handles us-gaap patterns today, so all IFRS concepts return None.

Missing balance sheet concepts: You're right that it's related to how AZN structures these line items. The specific issue is that the 5 concepts you listed (CurrentAssets, RightofuseAssets, NoncurrentDerivativeFinancialAssets, NoncurrentAssets, Assets) come from the standard IFRS taxonomy, which we don't fetch or parse. The facts for these concepts are in memory (they were read from the XBRL instance document), but the presentation tree nodes that reference them end up with a null element name — so the lookup silently fails and the values never appear.

Both issues are tracked in our broader IFRS support effort and will be addressed in phases. The view="standard" fixes (your case) are Phase 1 and don't require the full IFRS stitching work.