2 Define Matching Logic & Keys
Dataset A Key:
Dataset B Key:
+ Key 2 — A:
+ Key 2 — B:
+ Key 3 — A:
+ Key 3 — B:
e.g. match on Order ID + Line Number together. Columns are concatenated into one key before matching.
Dataset A Value:
Dataset B Value:
Dataset A Date:
Dataset B Date:
A Date Format:
B Date Format:
01/02/2026 is ambiguous (Jan 2 vs Feb 1) — set the format explicitly if Auto-detect can't confirm it from your data.
Verify 1 — A:
Verify 1 — B:
Verify 2 — A:
Verify 2 — B:
Checked only on clean 1-to-1 matches (exact/format/fuzzy, no duplicate keys) — flags a warning badge, doesn't change MATCH/MISMATCH status itself.
Same units as your Value column. Rounding slack still counted as MATCH.
Flags date gaps; buckets records for batch/split search. Needs both date columns.
Format/fuzzy/split-payment/batch matching for unmatched records.
e.g. INV12345 vs INV12346 can still match on Loose — check the confidence % shown.
3 Reconciliation Results
Total Keys Scanned
0
Reconciled
0
Value Mismatches
0
Orphan Records
0
≈ FORMAT = matched after normalizing case/punctuation
~ FUZZY = matched via approximate text similarity
Σ SPLIT = one record = sum of several on the other side
⊞ BATCH = many records on each side reconciled together (date-windowed)
⏱ TIME = same date & amount, no shared ID
⚑ VERIFY = a matched pair disagrees on an additional column you asked to check
⚠ Nd apart = IDs agree but dates differ more than your window
×N = duplicate keys merged from N rows
| Matching ID Key | Dataset A Value | Dataset B Value | Difference | Status |
|---|