Understanding reconciliation results

Convergence score

The convergence score is the single most important metric. It shows what percentage of records matched perfectly across both files, relative to all unique records combined.

Row counters

The four counters cover all rows across both files:

Numeric column statistics

For columns compared with a numeric condition (EQ, GT, GTE, LT, LTE), Reconcio computes aggregate statistics across matched rows where values differ. Reported separately for integer and float columns.

Join key statistics

The quality of a reconciliation depends directly on the join keys. These stats help identify duplicates, nulls, and low-effectiveness keys.

Column quality & detailed stats

Recommendations

Reconcio automatically generates recommendations based on the results. Each has a type and a details map with context-specific data.

Diff file

After a completed reconciliation you can download a diff file — a CSV containing every row that did not fully match. This is the file you send to whoever needs to investigate discrepancies.

The diff file does not include matched rows (rows where all condition checks passed). It contains only rows with at least one problem: a failed condition, or a row missing entirely from one file.

System fields — present in every row

Every row in the diff file contains three system fields that summarise the overall status of that record:

Value fields — original values from each file

For every non-join-key column configured in the mapping, the diff contains two value fields — one from each file. The naming pattern is:

For example, if your mapping includes a column named Amount, the diff will contain columns file_a_Amount and file_b_Amount. Compare them side by side to see exactly what differs.

Join key columns appear without a prefix — just the column name as configured in File A. They are always non-null since the join key must exist to create a diff row.

Check fields — per-condition result

For every non-join-key column, the diff also contains a check field named after the condition that was applied. The naming pattern is:

Complete example

Suppose you configured a reconciliation with:

• Join key: InvoiceId
• EQ condition on: Amount, Currency
• GTE condition on: Quantity

The diff file will have these columns in this order:

# System fields (always first)
all_rule_checks      UInt8     — overall pass/fail for this record
absent_in_file_a     UInt8     — 1 if record missing from File A
absent_in_file_b     UInt8     — 1 if record missing from File B

# Join key column (no prefix, always non-null)
InvoiceId            String    — the join key value

# Value fields for Amount (EQ condition)
file_a_Amount        Nullable(Float) — value from File A
file_b_Amount        Nullable(Float) — value from File B

# Value fields for Currency (EQ condition)
file_a_Currency      Nullable(String) — value from File A
file_b_Currency      Nullable(String) — value from File B

# Value fields for Quantity (GTE condition)
file_a_Quantity      Nullable(Int)   — value from File A
file_b_Quantity      Nullable(Int)   — value from File B

# Check fields (condition result per column)
eq_Amount            UInt8 — 1=pass, 0=fail or absent
eq_Currency          UInt8 — 1=pass, 0=fail or absent
gte_Quantity         UInt8 — 1=pass, 0=fail or absent

Sample data rows:

all_rule_checks,absent_in_file_a,absent_in_file_b,InvoiceId,file_a_Amount,file_b_Amount,file_a_Currency,file_b_Currency,file_a_Quantity,file_b_Quantity,eq_Amount,eq_Currency,gte_Quantity
0,0,0,INV-003,1000.00,1001.50,USD,USD,5,5,0,1,1  ← Amount mismatch
0,0,0,INV-007,500.00,500.00,EUR,GBP,3,3,1,0,1   ← Currency mismatch
0,1,0,INV-089,NULL,750.00,NULL,USD,NULL,2,0,0,0          ← missing in File A
0,0,1,INV-210,450.00,NULL,EUR,NULL,1,NULL,0,0,0           ← missing in File B
1,0,0,INV-015,200.00,200.00,USD,USD,4,3,1,1,1    ← all pass (gte: 4≥3 ✓)

Reconciliation statuses