Skip to content

Systems and records

Accounting chart mapping dataset: mapping charts of accounts by meaning

By SourceX Editorial · Updated

10 sections

Short answer

To map charts of accounts without losing meaning, map each source account by what actually posts to it: its type, normal balance, statement line, dimensions, typical entries and governing policy. Record the target, cardinality and any split rule, label the ambiguity, keep the reviewer's corrections, and prove the mapping by reconciling balances by statement line before and after. Kept this way, mapping decisions form a reusable dataset.

Value drivers

Accounting chart mapping

This page covers accounting chart mapping dataset: mapping charts of accounts by meaning.Illustrative

Key takeaways

  • Names and numbers are local conventions; type, normal balance, dimensions and actual postings carry the meaning.
  • ERP mapping fields store the outcome only, and a blank field can silently fall back to the same account number.
  • Record cardinality, split rules, rationale and approvals for every decision, including rejected first proposals.
  • A fixed set of ambiguity labels shows where judgment was needed and how it was resolved.
  • A mapping is proven only when statement-line totals reconcile before and after, period by period.
  • Policy differences, intercompany balances and currency translation go to an accountant, not a rule.

How do you map charts of accounts while preserving business meaning?

Map each source account by what actually posts to it, not by its name or number. For every account, establish its type, normal balance, statement placement, the transactions that land there and the policy behind them, then choose a target, record the cardinality (one to one, many to one or a split), label any ambiguity, and prove the result by reconciling balances by statement line before and after the mapping.

Charts of accounts are mapped constantly: when a subsidiary is consolidated into a group chart, when a company moves from one ERP to another, when an acquired business is folded in, and when a local statutory chart must be produced from an operating chart. In each case the risk is the same. Two accounts with similar names can hold different things, and two with different names can hold the same thing.

This page treats each mapping decision as a record worth keeping. Kept with its evidence and an ambiguity label, a mapping history becomes a reusable dataset: a reviewer can audit it, the next migration can reuse it, and a model can be trained or tested on the judgments it contains.

Why are account names and numbers weak evidence of what an account means?

Because names and numbering schemes are local conventions, and the meaning sits in configuration and usage that the label does not show. Mapping on labels alone is the most common way meaning gets lost.

Microsoft's Business Central documentation notes that businesses typically build a hierarchy into account codes, such as codes starting with 1 for assets, and that some regions mandate a standard chart. It also recommends dimensions, such as department or project, instead of separate accounts for each. One company may therefore carry a separate revenue account per region while another carries one revenue account with a region dimension. A name match between them misses the dimension entirely.

Account type carries meaning the name does not. Odoo's documentation says an account's type drives country-specific legal and financial reports, the rules for closing a fiscal year and the generation of opening entries. Two accounts called Prepaid Expenses can sit under different types in two systems and behave differently at year end.

Large ERPs separate the charts themselves. SAP describes an operating chart of accounts used for daily postings, an optional alternative chart for country-specific legal reporting with each operating account linkable to an alternative account, and a group chart whose group account numbers gather accounts under one financial statement item. A mapping between two companies may need to pass through one of these rather than go account to account.

What do ERP mapping fields record, and what do they leave out?

ERP mapping fields store the result of each decision, not the reasoning, so they are a starting point for a mapping dataset rather than the dataset itself.

In Business Central, each posting account in a business unit is mapped to an account in the consolidated company through the Consol. Debit Acc. and Consol. Credit Acc. fields; if they are left blank, the system assumes the same account numbers in the consolidated company. An account can be excluded from consolidation with a toggle, and each account carries a Consol. Translation Method, such as closing rate for balance sheet accounts or average rate for income statement accounts. A blank mapping that silently falls back to the same number is exactly the kind of decision a dataset should label, because it was never made by a person.

Odoo maps accounts across companies with a code per company on the account's Mapping tab. Its documentation gives the example of a subsidiary with five income accounts, from domestic product sales to freight and handling revenue, all mapped to a single parent income account. That rollup is correct for the parent's profit and loss and loses detail the subsidiary still needs, which is the trade-off a mapping record should state. A group that later needs product-line revenue back cannot recover it from the parent's books unless the mapping kept the source account or a dimension alongside the rollup. Odoo also allows accounts to be merged into shared accounts and warns that unmerging does not separate their change histories.

None of these fields says why the target was chosen, who approved it, what evidence was checked or how confident the reviewer was. That is what the decision dataset adds.

The mapping decision dataset, field by field

Keep one row per source account per mapping event, with the evidence attached. The field list below works for consolidation, migration and statutory mapping alike. Example entries are illustrative.

The mapping decision dataset, field by field
FieldWhat to recordIllustrative entryWhy it matters
Source accountEntity, chart, number, name, type, normal balance, posting or headingSub A, 2150 Accrued Liabilities, current liability, credit, postingType and balance carry meaning the name does not
Usage profileTop posting sources and typical entries in the sample periodMonthly payroll accrual and reversal; year-end bonus accrualShows what the account really holds
Dimensions in useDimension values that split the account's balanceDepartment, cost centerA dimension may need to become a separate target account
Target accountTarget chart, number, name, type and statement lineGroup 2100 Accrued PayrollThe decision itself
CardinalityOne to one, many to one, one to many split, or excludedOne to many splitSplits need an allocation rule
Split or allocation ruleHow a split is made and on what driverPayroll accruals to 2100, bonus accruals to 2110 by journal sourceWithout a rule a split cannot be reproduced
Ambiguity labelOne label from the scheme in the next sectionScope mismatchTells reviewers and models where judgment was needed
Rationale and evidenceWhy this target, citing journal samples, policy or prior mappingBonus accruals are reported separately in group policyLets a reviewer check the reasoning
ApprovalPreparer, reviewer, date, and any changed first proposalPreparer proposed 2100 only; reviewer splitCorrections are the most informative records
Effective period and statusPeriods the mapping applies to; active, superseded or reversedFrom the first period of the fiscal year; activeMappings change and old periods must stay reproducible

Which ambiguity labels should each mapping decision carry?

Use a small, fixed set of labels so reviewers apply them consistently. Each label names a reason a mapping needed judgment, and the resolution column says what evidence closes it.

Which ambiguity labels should each mapping decision carry?
LabelMeaningTypical evidenceResolution
CleanSame type, balance, content and statement lineJournal sample matches target definitionMap one to one
RollupSeveral source accounts collapse into one targetSource accounts share type and statement lineMap many to one; keep source detail in a dimension if needed
Split requiredOne source account holds content that belongs in two or more targetsJournal sources or descriptions differ by contentDefine an allocation rule or restructure the source
Scope mismatchNames match but contents differ in coverageSample entries outside the target's definitionMap to a different target or split
Balance conflictSource and target normal balances differ, as with contra accountsAccount type and sign of typical entriesMap to a contra target or flag for presentation
Policy differenceSource follows a different accounting policy than the target entityPolicy documents for both entitiesAccountant decides; may need an adjustment entry
Translation dependentMapping or value depends on currency translation methodTranslation method on the account and entity currencyRecord the method; review with consolidation lead
No targetNothing in the target chart fitsReviewer confirmationCreate a target account or exclude with approval

Worked example (illustrative): folding a subsidiary's accounts into a group chart

A hypothetical services subsidiary is consolidated into a group chart. Most of its 140 posting accounts map cleanly. Five decisions show where meaning would have been lost on names alone.

The exception in this example is a clearing account the subsidiary uses for card settlements. It has no meaningful group equivalent and should carry a zero balance at period end. The team labeled it no target, excluded it from consolidation with approval, and added a check that its period-end balance is zero; a non-zero balance blocks the close and sends the account to review instead of being mapped somewhere convenient.

Notice that two of the five decisions could not have been made from the chart alone. The deferred revenue scope problem only appeared when someone read a sample of the entries posting to the account, and the bonus split only appeared when the journal sources were listed. That is why the usage profile is a required field rather than an optional note.

  • Five revenue accounts by product line map to one group revenue account: labeled rollup, with product line kept as a dimension so subsidiary reporting survives.
  • Accrued liabilities hold both payroll and bonus accruals: labeled split required, with an allocation rule by journal source.
  • An account named Deferred Revenue also holds customer deposits for work not yet scheduled: labeled scope mismatch; deposits map to a separate group liability.
  • Allowance for doubtful accounts carries a credit balance under assets: labeled balance conflict and mapped to the group contra-asset account, not to receivables.
  • Development costs are expensed under the subsidiary's policy but capitalized under group policy: labeled policy difference and sent to the group controller, with an adjustment rather than a remap.

Can the reviewer reconcile the mapping to documents, authority and the final financial state?

A mapping is only proven when balances reconcile. The test is mechanical, and every failure points to a specific decision row.

Normal balance is worth checking explicitly because it is part of meaning, not presentation. Reporting standards encode it: in XBRL 2.1, monetary concepts can carry a balance attribute of debit or credit, and every concept carries a period type of instant or duration. A source account mapped to a target, or a reporting concept, with the opposite normal balance should either be a labeled contra account or be wrong.

Reconciliation proves the totals; it does not prove authority. Each non-clean decision should also point to the document that justified it, such as the group accounting manual, a policy memo or the controller's approval, and to the journal entries sampled when the usage profile was built. A reviewer who can open those in a few minutes can confirm the decision; one who has to ask the preparer cannot.

  • Trial balance totals by statement line agree before and after mapping for the same periods and entities.
  • Every source posting account with activity in the period has exactly one active mapping row or an approved exclusion.
  • Split accounts reproduce from their allocation rule to the same totals.
  • Normal balance and sign agree between source and target, or the row is labeled balance conflict.
  • Accounts that should be zero at period end, such as clearing and suspense accounts, are zero.
  • Each non-clean row has a named reviewer and an approval date before the period is reported.
  • Superseded mappings still reproduce the balances reported for their periods.

What makes a mapping history useful as AI training or evaluation data?

A mapping history is useful when each record pairs the inputs a person had with the decision they reached and the evidence that it was right. The inputs are both charts with their types and balances, a sample of postings per source account and the relevant policies; the output is the target, the cardinality, the label and the rationale; the proof is the reconciliation.

At a glance

What makes a mapping history useful as AI training or evaluation data

The most informative records are the ones where a first proposal was changed by a reviewer, because they show which cues a name-based guess misses. Keep the rejected proposal alongside the final decision rather than overwriting it. Split evaluation sets by entity or by mapping event, not by random rows, so a model is not tested on accounts from a chart it has already seen.

Leave out what the task does not need. Posting samples often carry customer and employee names in descriptions; replace them with role labels before the records leave the company, and keep amounts as they are only if the reconciliation test needs them.

  • Proposal task: given both charts and a posting sample, propose a target and cardinality for each source account.
  • Labeling task: given a proposed mapping, assign the ambiguity label and say what evidence would resolve it.
  • Review task: given a preparer's mapping, find the rows a reviewer changed and explain why.
  • Reconciliation task: given a mapping and two trial balances, identify which rows cause a difference.

When does a mapping need an accountant rather than a rule?

Stop mapping and escalate when the difference between source and target is an accounting judgment rather than a classification. A rule or a model can propose a target; it cannot decide which policy governs.

The usual triggers are accounts governed by different accounting policies in the two entities, intercompany accounts that must eliminate, equity and retained earnings in entities with different functional currencies, accounts whose translation method is not obvious, and any mapping that would change a reported figure for a closed period. In those cases the record should carry the policy difference or translation dependent label, the name of the person who decided, and the authority they relied on.

Escalation is also the right answer when the evidence runs out. If nobody can say what an old account was used for and its postings are too few to tell, map it provisionally, label it, and put it on the reviewer's list rather than guessing with confidence. A labeled guess is honest data; an unlabeled one teaches the next person, and any model trained on the history, that the guess was a fact.

Licensing chart mapping histories

Mapping histories with approvals, corrections and reconciliations are operational finance records like any other. For finance teams that license such histories to AI developers, SourceX covers the path from sourcing and rights review to contracting, preparation, delivery and payment. Its general guide for finance histories is two or more years recorded consistently in one system and tens of thousands of records, with exceptions for rarer work, and each case is assessed on its own. A controller can learn whether the company qualifies through a short assessment without sending ledgers or mapping files.

Frequently asked questions

What is a chart of accounts mapping?

It is the set of decisions that assigns each account in one chart to an account in another, for consolidation, a system migration or statutory reporting. A complete mapping also records splits, exclusions and the reasons behind each choice.

Can chart of accounts mapping be automated?

Matching on names, numbers and account types can propose most clean mappings. Splits, scope mismatches, contra accounts and policy differences need someone who can read the postings and the policies, so automated proposals should be reviewed and the corrections kept.

What happens in Business Central if consolidation accounts are left blank?

Microsoft's documentation says that if the Consol. Debit Acc. and Consol. Credit Acc. fields are empty, Business Central assumes the account maps to the same account number in the consolidated company. That fallback should be reviewed, not assumed.

How do you check a chart of accounts mapping is correct?

Reconcile trial balance totals by statement line before and after mapping for the same periods, confirm every active account has one mapping or an approved exclusion, and check normal balances agree or are labeled as contra accounts.

What is an ambiguity label in a mapping dataset?

It is a short code on each mapping decision that says why judgment was needed, such as rollup, split required, scope mismatch or policy difference. It tells reviewers where to look and lets a model be evaluated separately on easy and hard cases.

Sources

  • Business Central maps each business unit posting account to a consolidated account through Consol. Debit Acc. and Consol. Credit Acc., assumes the same account when they are blank, allows exclusion from consolidation and sets a translation method per account. Source
  • Business Central documentation notes hierarchical account codes, regional standard charts, and dimensions as an alternative to separate accounts. Source
  • Odoo maps accounts across companies with per-company codes, gives an example of five subsidiary income accounts mapped to one parent account, and warns that unmerging accounts does not separate their histories. Source
  • Odoo says an account's type drives country-specific reports, fiscal year closing rules and opening entries. Source
  • SAP S/4HANA Cloud distinguishes the operating chart of accounts, an optional country-specific alternative chart and the group chart. Source
  • SAP group account numbers group G/L accounts so they are displayed under one financial statement item. Source
  • XBRL 2.1 allows a debit or credit balance attribute on monetary item concepts and requires a period type of instant or duration. Source

See if your company qualifies

A short company assessment. No data uploads are needed.

See if your company qualifies