Source Document → Register Movements → Report: Finding the First Divergence in BAS/1C
An accounting discrepancy is often discovered at the end of a long data path: a balance is wrong, a report line is missing, VAT appears in the wrong period, or a counterparty debt does not match the supporting documents.
That final symptom is useful, but it is rarely the best place to begin debugging.
In BAS/1C, a business transaction normally passes through several distinct states:
source evidence
→ document fields
→ validation and posting logic
→ register movements
→ accumulated balances
→ report query
→ report presentation
A reliable investigation should locate the first divergence between expected and actual data. Everything after that point is usually a downstream consequence.
This article presents a repeatable protocol for developers, consultants, and senior users who need to distinguish among:
- incorrect source data;
- incorrect document attributes;
- configuration or extension defects;
- unexpected register movements;
- period, dimension, or filter mismatches;
- report-query defects;
- presentation-only problems.
The goal is not merely to make a report “look right.” The goal is to identify the earliest incorrect state, apply the smallest justified correction, and prove that the result remains correct throughout the entire chain.
Why report-first debugging often fails
A report is an interpretation of stored accounting state. It may contain:
- account and subaccount filters;
- organization and period restrictions;
- grouping by analytical dimensions;
- virtual-table parameters;
- access restrictions;
- calculated fields;
- presentation rules;
- suppressed zero values;
- currency or quantity representations.
Because of this, a wrong report value does not automatically prove that the posting logic is wrong. Conversely, a correct-looking report does not prove that the underlying movements are correct. Two errors can compensate for each other and produce a plausible total.
Consider a simple example:
- one invoice for 12,000 was omitted;
- another invoice was duplicated for 12,000.
The closing balance may still match the expected total, although the transaction history is wrong. A total-only check would pass. A document-to-movement reconciliation would fail.
A good diagnostic method therefore treats the report as the entry point to the investigation, not as the authoritative description of the cause.
Define the control target before opening the debugger
Before changing code, reposting documents, or rebuilding totals, write down the exact control target.
At minimum, capture:
| Field | Example |
|---|---|
| Information base | Production accounting base |
| Organization | Example Trade LLC |
| Period | 2026-10-01 through 2026-10-31 |
| Report | Trial balance by account |
| Account | Supplier settlements |
| Analytical dimensions | Supplier, contract, settlement document |
| Expected result | Credit balance 48,000 |
| Actual result | Credit balance 36,000 |
| Difference | 12,000 |
| Evidence | Supplier statement and accepted invoice |
| First known occurrence | After posting invoice PI-1045 |
This turns “the report is wrong” into a falsifiable statement.
It also prevents scope drift. During an investigation, it is easy to notice unrelated inconsistencies and begin fixing them. The control target keeps the team focused on one discrepancy until its complete evidence chain is understood.
Step 1: Validate the source evidence
The source document is not necessarily a PDF or a piece of paper. It may be:
- an invoice;
- a bank statement;
- a warehouse receipt;
- a tax document;
- a signed act;
- a CRM order;
- an API payload;
- an import file;
- an approved internal calculation.
Determine what evidence authorizes the transaction and which values it establishes.
For a supplier invoice, record at least:
- external document number;
- document date;
- supplier;
- contract or settlement basis;
- organization;
- currency;
- amount before tax;
- tax amount;
- total amount;
- goods or services received;
- expected accounting period.
Do not assume that the source is correct merely because it exists. Check for amended invoices, credit notes, duplicates, cancelled documents, partial acceptance, and differences between the transaction date and the accounting date.
At this stage, the question is:
What should be recorded if the system accurately reflects the approved business event?
That expected state becomes the baseline for every later comparison.
Step 2: Compare source evidence with document fields
Open the BAS/1C document that should represent the transaction. Compare its significant attributes with the baseline.
Typical fields include:
- organization;
- counterparty;
- contract;
- settlement document;
- operation type;
- warehouse;
- business unit;
- document date;
- posting date;
- currency and exchange rate;
- item, service, or expense classification;
- tax rate and tax treatment;
- accounts and analytical dimensions;
- amount, quantity, and price.
A frequent mistake is to compare only visible header fields. Posting logic may depend on values stored elsewhere:
- an account in an item or counterparty card;
- a functional option;
- an accounting-policy register;
- an agreement type;
- a tax-purpose attribute;
- an organization-specific setting;
- a default filled by an extension;
- a hidden attribute populated during import.
Record both the expected and actual values. If the first mismatch appears here, debugging the report query is premature.
A useful evidence table
Maintain a compact trace table as the investigation proceeds:
| Stage | Expected | Actual | Match? | Evidence |
|---|---|---|---|---|
| Source invoice | Supplier A, 12,000 | Supplier A, 12,000 | Yes | Accepted invoice |
| BAS/1C document | Supplier A, Contract 2026-01 | Supplier A duplicate, Contract 2026-01 | No | Document fields |
| Posting | Credit Supplier A, 12,000 | Credit Supplier A duplicate, 12,000 | No | Register movements |
| Trial balance | Supplier A credit 48,000 | Supplier A credit 36,000 | No | Report drill-down |
The table makes the first divergence visible: the document selected a duplicate counterparty record. Incorrect movements and the report discrepancy are consequences.
Step 3: Confirm document lifecycle and posting state
Before inspecting algorithms, confirm that the document participated in accounting at all.
Check:
- whether it is posted;
- whether posting succeeded without rollback;
- whether it was reposted after attributes changed;
- whether it was marked for deletion;
- whether another document reversed it;
- whether the operation belongs to the requested organization and period;
- whether an exchange later overwrote or cancelled it;
- whether background processing was required;
- whether the document has active movements.
“Visible in the journal” does not mean “included in balances.” A document can exist without producing current movements.
Also distinguish document time from registration time. A document dated 31 October may have been entered in November, reposted with another timestamp, or intentionally reflected in a different period according to the configuration’s rules.
Step 4: Inspect register movements as an accounting contract
Register movements are the most useful boundary between business logic and reporting.
Treat them as a contract:
- the document provides input;
- the posting algorithm applies configuration rules;
- the registers receive a normalized accounting result;
- reports consume that stored result.
For each relevant movement, capture:
- registrar;
- period;
- activity flag;
- organization;
- account or resource;
- debit and credit sides where applicable;
- analytical dimensions;
- amount;
- quantity;
- currency amount;
- tax attributes;
- settlement document;
- any configuration-specific dimensions.
Then reconstruct the expected movement independently from the source evidence and the documented accounting rule.
Do not begin by asking whether the code “looks correct.” Ask whether the movement represents the business event correctly.
Classify movement defects
Most movement defects fall into a small number of categories:
- Omission — An expected movement does not exist.
- Duplication — The same economic effect is recorded more than once.
- Misclassification — The amount is correct, but the account, register, or operation type is wrong.
- Analytical mismatch — The total is present, but it is attached to the wrong counterparty, contract, item, warehouse, project, or settlement document.
- Timing error — The movement belongs to the wrong date or period.
- Direction error — Debit and credit, receipt and expense, or increase and decrease are reversed.
- Measurement error — Amount, quantity, rate, currency amount, or tax is wrong.
- Stale movement — The document changed, but old movements were not replaced as expected.
This classification helps choose the next diagnostic step and prevents vague conclusions such as “posting is broken.”
For a conceptual refresher on how a business document produces postings and how those postings later appear in reports, see Postings: from the document to the report.
Step 5: Locate the responsible rule
If document fields are correct but movements are wrong, inspect the logic that transforms the document into movements.
Possible sources include:
- standard posting procedures;
- object or manager modules;
- common modules;
- event subscriptions;
- extensions;
- exchange handlers;
- background jobs;
- pre-posting validation;
- account-selection rules;
- organization policy settings;
- effective-dated information registers.
Use the smallest reproducible scenario. Avoid debugging a large production document with dozens of lines if a one-line copy reproduces the issue.
A controlled comparison is especially useful:
| Variable | Test A | Test B |
|---|---|---|
| Platform build | Same | Same |
| Configuration release | Same | Same |
| Extension set | Enabled | Disabled |
| Information base | Production copy | Clean test base |
| Input values | Equivalent | Equivalent |
| Result | Wrong account | Correct account |
This does not immediately prove that the extension is defective, but it narrows the investigation to a meaningful difference.
Change one variable at a time. If the platform build, configuration release, permissions, data, and settings all differ, the comparison provides weak evidence.
Step 6: Reconcile movements with accumulated balances
Correct individual movements do not always guarantee correct balances. Check the layer between movements and the report.
Potential causes include:
- an incorrect period boundary;
- inactive or reversed movements;
- a different analytical dimension;
- totals not recalculated after an abnormal operation;
- later documents changing the balance;
- exchange duplication;
- currency versus accounting-currency confusion;
- report use of another register;
- sequence or closing operations;
- access restrictions excluding part of the data.
Calculate the expected balance from the relevant movements, not from the final report presentation.
For a simple balance:
expected closing balance
= opening balance
+ relevant increases
- relevant decreases
In accounting registers, respect debit/credit semantics and the configuration’s treatment of active/passive accounts. In accumulation registers, respect the movement type and all dimensions used by the virtual table.
If the stored movements produce the correct balance, the investigation moves to the report layer. If they do not, remain in the movement, period, and totals layers.
Step 7: Inspect the report query and parameters
Only after the underlying state is understood should you investigate the report.
Capture the effective parameters, not merely what the form appears to show:
- start and end dates;
- organization;
- account;
- dimensions;
- hierarchy settings;
- comparison type;
- inclusion or exclusion filters;
- currency;
- periodicity;
- data-composition settings;
- access restrictions;
- saved user variants.
Then inspect which data source the report actually queries.
Common report defects include:
- using a different register than expected;
- omitting an analytical dimension in a join;
- applying a filter after aggregation instead of before it;
- using an exclusive end date unintentionally;
- joining one-to-many data and duplicating amounts;
- converting null values incorrectly;
- suppressing meaningful zero or negative rows;
- selecting turnover where balance is required;
- mixing organization-specific and global data;
- using presentation values as join keys;
- relying on a saved user setting that is no longer valid.
A report-query defect is established when the stored accounting state is correct but the query or its parameters transform it incorrectly.
Worked example: the missing 12,000
Assume the October supplier trial balance is 12,000 lower than the supplier statement.
Observation
The report shows:
Supplier A
Opening credit: 24,000
October credit turnover: 24,000
Payments: 12,000
Closing credit: 36,000
The expected closing credit is 48,000.
Source evidence
There are two accepted October invoices for 12,000 each. Both are valid and unpaid at month-end.
Expected October credit turnover: 36,000, including another 12,000 transaction already recorded earlier in the month.
Document comparison
Both BAS/1C documents exist and are posted.
The first uses Supplier A.
The second uses another catalog entry with the same visible name and tax identifier, created during an import. Its internal reference is different.
Register movements
Both documents generated correct amounts and accounts. However, the second invoice’s settlement movement is attached to Supplier A (duplicate).
Balance reconciliation
The combined supplier balance across both catalog references is 48,000. The report row for the canonical Supplier A correctly shows only 36,000.
First divergence
The first divergence occurs in the document field that references the counterparty. It is not a platform defect, a posting-algorithm defect, a totals defect, or a report-query defect.
Minimal correction
Use the authorized data-cleanup procedure:
- replace the duplicate reference in the affected document;
- repost the document;
- confirm that the old analytical movement is removed;
- confirm that the new movement uses the canonical counterparty;
- regenerate the report with identical parameters;
- reconcile the combined supplier balance;
- implement duplicate prevention or detection.
Changing the report to merge counterparties by tax identifier would conceal a master-data defect and could incorrectly combine legally distinct records.
Anti-patterns to avoid
“Repost everything”
Mass reposting can change historical results, trigger integrations, hide the original evidence, and make the discrepancy harder to reproduce. Repost only when the investigation identifies a justified document-level correction.
“Rebuild totals first”
A totals rebuild is appropriate only when evidence indicates totals corruption. It should not be used as a generic remedy for incorrect dimensions, dates, or movements.
“Fix the report until the number matches”
Hard-coded exceptions, additional joins, or aggregation changes can make one control value look correct while damaging other periods and dimensions.
“Edit register records directly”
Direct register edits break the relationship between registrar and movement, weaken the audit trail, and may be overwritten on the next posting. Correct the authoritative source through supported mechanisms.
“Compare environments by name”
“Test” and “production” may differ in platform builds, configuration releases, extensions, policies, permissions, or historical data. Capture exact identifiers before using one environment as evidence for another.
“Stop after the target number changes”
A correction is not proven merely because one report value now matches. Verify the entire affected chain and confirm that unrelated balances remain stable.
Validation checklist
After applying the smallest justified correction, verify:
- [ ] The source evidence and document fields match.
- [ ] The document lifecycle and posting state are correct.
- [ ] Expected movements exist exactly once.
- [ ] Accounts, directions, periods, amounts, and dimensions are correct.
- [ ] Obsolete movements were removed.
- [ ] Relevant balances reconcile with the movement set.
- [ ] The report uses the intended period, organization, and filters.
- [ ] Drill-down reaches the corrected registrar.
- [ ] The original discrepancy is removed.
- [ ] Unrelated accounts, counterparties, periods, and organizations did not change.
- [ ] The correction is documented with before-and-after evidence.
- [ ] A preventive control is assigned where appropriate.
Preserve an evidence packet
For defects that may recur, retain a compact evidence packet:
- anonymized source document;
- document identifier and significant attributes;
- platform and configuration versions;
- extension list;
- relevant policy and settings;
- before/after register movements;
- report parameters;
- before/after report extracts;
- root-cause classification;
- exact correction;
- regression checks.
This turns an isolated incident into a reusable diagnostic case and makes future testing more precise.
Conclusion
A BAS/1C discrepancy should be investigated as a sequence of state transitions, not as an isolated report problem.
Start with the approved business evidence. Compare it with document fields. Confirm posting state. Treat register movements as the accounting contract. Reconcile movements with balances. Inspect the report only after the stored state is understood.
The decisive question is:
At which stage does the actual state first differ from the expected state?
Finding that first divergence produces smaller corrections, stronger evidence, safer regression tests, and fewer “fixes” that merely move the discrepancy somewhere else.
Disclosure: AI assisted with drafting and editing; the final structure, technical claims, examples, and publication were reviewed by a human.
All rights reserved