Mapping
These columns represent the same business concept. Mapping does not make a field part of the key or a comparison rule.
MatchDock help
How MatchDock turns two exports into reviewable exceptions.
Map equivalent columns, define the business key that identifies a record, apply explicit comparison rules, and review every uncertain outcome instead of accepting an invented match.
MatchDock User Manual
Complete guidance on files, rules, Results, saved workflows and evidence exports.
Composite business key
Three fields identify one payroll record
| Source | Employee ID | Wage Type | Effective Date |
|---|---|---|---|
| A | 000123 | 1000 | 2026-08-01 |
| B | 123 | 1000 | 2026-08-02 |
Complete workflow
Five steps from files to evidence
01
Select sources
Choose one CSV or standard XLSX file for Source A and Source B.
02
Map columns
Confirm which columns represent the same business concept.
03
Define key
Choose one to five mapped fields that identify a business record.
04
Set comparison rules
Choose compare fields, normalization, number and date formats, and any permitted tolerances.
05
Review and export
Run locally, begin with the exceptions, inspect each explanation, and export the evidence.
Start with the source structure
Confirm what MatchDock detected before defining how records should match.
- CSV
- MatchDock supports comma, semicolon, and tab-delimited files, including valid quoted fields and embedded line breaks.
- Standard XLSX
- Choose one worksheet from a standard .xlsx workbook. The first visible sheet is selected by default and can be changed.
- Header row
- Row 1 is the default header. Change it when an export contains report titles or parameters above the table.
- Local preview
- Confirm the detected columns and a limited row preview before mapping.
See current limitations for file, row, and column caps.
Mapping is not matching
Keep column equivalence, record identity, and value comparison as three explicit decisions.
Bank reconciliation is available in Templates / starting points. Confirm shared transaction references and exact amount, date and currency comparisons. The template keeps signs unchanged. In Rules, NUMBER compare fields offer Keep sign (KEEP) or Negate sign (NEGATE) independently for Source A/B. Polarity is applied after parsing and before exact comparison and tolerances; raw values remain unchanged. Keys and TEXT/DATE fields cannot use polarity. Separate debit/credit columns still require manual preparation; this is not accounting.
Key
These fields identify which record in Source A corresponds to which record in Source B.
Compare fields
These fields are checked only after MatchDock has identified a unique pair.
Composite example
Keys answer identity
A key may use one to five mapped fields. For example: Employee ID + Wage Type + Effective Date.
Amount or Cost Center can then be compared without silently becoming part of the key.
Composite business key
Three fields identify one payroll record
| Source | Employee ID | Wage Type | Effective Date |
|---|---|---|---|
| A | 000123 | 1000 | 2026-08-01 |
| B | 123 | 1000 | 2026-08-02 |
What each result means
Each category describes a confirmed reconciliation state, including records that MatchDock deliberately leaves unresolved.
- Differences
- The key identifies one row on each side, but at least one compared field differs beyond its rule.
- Only A
- A valid unique key exists in Source A with no corresponding row in Source B.
- Only B
- A valid unique key exists in Source B with no corresponding row in Source A.
- Duplicates
- The key occurs more than once on either side, or a tolerant date match is not mutually unique.
- Invalid keys
- A required key component is empty or cannot be interpreted using the selected rule.
- Matched
- All compared fields agree exactly.
- Within tolerance
- No field mismatches, but at least one approved tolerance was used.
Tolerances stay explicit
Numeric comparison
NUMBER uses exact decimal arithmetic: up to 38 significant digits and 18 fractional places, with absolute values below 10^38 and input tokens up to 1,024 UTF-16 units. Unsupported values are rejected, never rounded. Text scientific notation is not supported. Number compare fields support exact matching, absolute tolerance, or percentage tolerance. Percentage is measured against the larger absolute value: 1 means 1%. When both tolerances are enabled, either can accept a difference. Percentage does not apply when only one value is zero or the signs differ; an explicit absolute tolerance can still apply.
Typed spreadsheet numbers are reconciled from the value delivered by the reader, subject to safe-integer and 15-significant-digit limits. Digits already lost by Excel or the reader cannot be recovered. CSV and textual XLSX cells provide the strongest exact-decimal path; use TEXT for identifiers whose exact spelling matters. No monetary rounding, currency conversion, aggregation, or certified accounting is implied. Exact normalized numeric evidence is exported as text.
Date keys
At most one date field in a key may use ±N days. A tolerant date pair is accepted only when it is unique in both directions; otherwise it remains ambiguous.
Additional empty values
In Rules, configure additional empty tokens for each field and Source A/B, one token per line. Matching is case-sensitive and ignores outer spaces; internal spaces and quotes are literal. No tokens such as NULL or N/A are enabled by default. Blank cells are always empty. A key value matching a token becomes Invalid Key; two empty compare values agree exactly.
Tokens apply to textual cell values before number/date parsing, without converting typed cells to text. For example, token “0” can empty a CSV string or XLSX text “0”, but an XLSX numeric zero remains numeric. Typed dates and booleans are also unchanged. Review tokens carefully: they can hide legitimate values. Results and Excel exports preserve the original values and the effective rules.
Export the evidence
Export Excel creates a local workbook with Summary, Differences, Only Source A, Only Source B, Duplicates, Invalid Keys, Matched, and Rules sheets. The Rules sheet records the configuration used for the result. Filenames are not required and are not included. User-derived strings are protected against accidental formula execution.
Excel export is available when the result fits the physical worksheet limits. Evidence Package is a separate, explicit ZIP download with a manifest and eight evidence CSVs preserving raw values, normalized values, explanations and Rules. It is recommended for very large evidence and remains available when XLSX cannot represent the result. MatchDock never silently truncates or switches export formats.
Pro also includes Professional Report: a whole-run preview, a separate three-sheet XLSX (Summary, Configuration, Exceptions) and self-contained printable HTML. Filenames and bounded exception samples are Off by default. Search and filters do not change its facts. Use your browser’s Print / Save as PDF command for HTML. Reports support human review; they do not certify or approve the data and do not replace complete evidence exports. Free keeps its existing exports.
In the same Pro dialog, Client-ready mode adds a local native PDF, English or Spanish report content, limited branding, an optional PNG/JPEG logo and explicit local presentation profiles. Financial summary is Off until you select one NUMBER field and declare its common currency or unit. TEXT currency validation is optional; conflicting or missing validated units prevent totals. No currency conversion or monetary rounding is performed. The two charts show counts, not financial impact. Report ID and the local SHA-256 fingerprint relate the report formats; they are technical references, not a signature or certification. PDF font and page guards can reject unsupported text or unusually large layouts; HTML and XLSX remain separate explicit choices. No report data or profile is uploaded.
The application language selector switches between English and Spanish and saves its choice in this browser. It does not change report language or workflow data. Free works without an account. Anonymous Get Pro opens the existing email-code sign-in or account-creation flow and returns to the Pro journey; Sign in remains ordinary account access. Authenticated Free accounts see Upgrade to Pro and can choose €15 monthly or €144 annually (equivalent to €12/month and 20% less than twelve monthly payments). Paddle-backed Pro accounts see Manage subscription; courtesy Pro remains active without purchase or Paddle-management actions. Opening checkout alone does not grant Pro; the account service must confirm access. Signing out returns the header to its anonymous actions.
MatchDock Pro is licensed for use by one person. You may use it on multiple devices you control. Sharing an account with other people is not permitted. Each person needs their own account and Pro subscription.
MatchDock allows one authenticated session per account, for Free and Pro. A new sign-in automatically replaces the previous session. Tabs sharing the same session can be used together. MatchDock does not require manual session renewal or impose an inactivity timer. Already admitted work may finish; your local files, results and saved workflows are preserved. Anonymous Free use and the fixed demo remain available.
SummaryDifferencesOnly AOnly BDuplicatesInvalid KeysMatchedRules
Advanced Matching
Advanced is available while the service is active. Pro is required for your own files; Free can explore the fixed synthetic demo. Legacy, Bank and sign normalization remain separate.
Configure the scope
In Rules, explicitly enable Advanced, select complete groups or bounded subsets and the permitted cardinalities, then configure shared context, units and every comparison. Review the configuration before Run. A group compares aggregates; it does not invent row-to-row allocations.
Read both coverages
Results shows coverage and reconciled rows separately for Source A and Source B. Expand groups to inspect every member, original and effective values, totals, differences and checks. One grouped unit can own several rows.
Resolve uncertainty explicitly
Alternative subsets stay unresolved when uniqueness is not proved. N:M and fuzzy candidates always require human confirmation, including a similarity score of 100. Fuzzy needs an explicit algorithm and threshold; it never changes NUMBER or DATE values.
Review and keep a local journal
Accept, reject, replace or undo through the review queue. Confirmed differences remain financial differences. Export/import the compatible local journal to replay decisions; a cloud workflow stores reusable configuration, not files, row memberships, decisions or reports.
Understand limits and export status
Incomplete means part of the permitted search was not examined. Narrow the context rather than assuming there are no alternatives. CSV/XLSX input maxima do not certify Advanced capacity. Cancelled or failed runs and exports are not completed results.
Save complete evidence
Advanced Detailed XLSX and Package4 preserve members and the frozen review revision. XLSX may continue a logical category across physical sheets. Wait for successful file closure. The 192 MiB output allowance is not a guaranteed portable download; memory, work and destination limits also apply.
New Advanced operations and writes require current service authorization. If availability is paused or disabled, existing compatible workflows remain readable. Check account status and connectivity, or contact support.
Read the complete V2 manualCommon problems
Resolve these checks before changing the reconciliation rules.
- Unsupported file
- Use CSV or a standard .xlsx workbook. Legacy .xls files are not supported.
- Password-protected workbook
- Save an unencrypted standard .xlsx copy and try again.
- No data rows
- Check the selected sheet and header row.
- Missing mapping
- Every key or compare field needs a confirmed mapping.
- Ambiguous number or date
- Choose the source format explicitly instead of relying on a guess.
- Browser memory pressure
- Close other memory-intensive tabs or use a smaller source extract if available.
Start with a real workflow or the synthetic demo.
Match by a business key, keep ambiguous records visible, and review the exceptions before exporting the evidence.
Review current limitations