Physical Count & Recount Consistency Evidence

What does Physical Count & Recount Consistency Evidence do?

Turn authorized baseline, first-count and actual recount exports into exact-key evidence of agreement, disagreement and missing or ambiguous quantities. This batch tool compares the session, SKU, location, lot and unit you supply. It produces a report you can review before deciding what your warehouse process requires next.

A recount marked “done” does not supply its measured quantity. This report needs the actual second quantity and the explicit IDs of both passes. It neither collects counts nor posts inventory adjustments. It does not determine stock truth, valuation, shrinkage, theft or compliance.

Why use a count and recount report?

Use it when consolidating cycle-count exports from several counters, handing a discrepancy batch to an independent reviewer, or checking whether a recount changed the first observation. Run a new batch through the Apify API or an integration after each authorized export. A native warehouse system may already handle collection and approval; this tool provides a bounded supplied-data report for the handoff.

Different lot and unit keys remain separate. A kilogram baseline is never converted to a gram count, and an empty lot identifier is never joined to a tracked lot. Related keys appear as review references without becoming matches. Duplicate full-key rows within a pass remain ambiguous, including identical duplicates: no summing, averaging or last-row choice can hide double entry.

How to compare physical count passes

1. Export your baseline and actual count rows for the intended session. Use opaque authorized identifiers and remove contact details.
2. Assign a distinct rowId to every source row within each table. Include an explicit lotId; use an empty string only for your declared untracked lot scope.
3. Set distinct firstPassId and recountPassId. Choose cutOffEvidence and baselineCoverage explicitly.
4. Supply the two arrays, or remove the relevant array and supply its mapped CSV alternative. The input tab includes a synthetic example with both useful and problematic cases.
5. Run the Actor and download the per-key CSV, findings CSV, complete JSON or standalone HTML. Review unresolved keys and retained source references before drawing conclusions.

Input and cutoff requirements

Baseline rows contain rowId, sessionId, sku, locationId, lotId, unit, and quantity. Count rows add passId and use countedQuantity. Identifiers are exact case- and whitespace-sensitive strings; leading zeros are preserved. Quantities are nonnegative decimal strings with up to 24 integer and 12 fractional digits. Exponents, commas, numeric JSON values and negative quantities are rejected.

Each count row must represent one full key in one pass. Pre-consolidate legitimate separate counting areas under your own documented process; do not submit them as multiple full-key rows expecting this tool to sum them. A missing row is never zero. An explicit zero is a valid observation. Duplicate row IDs are invalid input, while distinct row IDs carrying a duplicate count key are retained as ambiguity.

comparable means you declare the baseline and selected passes comparable at the relevant cutoff. The Actor does not verify stock movements or this declaration. With unverified, raw recount-minus-first arithmetic remains available, but baseline differences are null and variance status stays unresolved. Equal pass quantities never prove inventory truth. partial baseline coverage prevents a whole-session consistency conclusion even when a particular supplied key agrees. Neither coverage option turns a missing baseline into zero.

CSV accepts BOM and quoted cells. Every required mapped header must be present exactly once; unknown headers, irregular rows and ambiguous mappings fail validation. Header-only CSV is explicitly empty evidence. Source references identify normalized rows such as countRows[0], not original physical line numbers when quoted multiline cells are present.

Limits are 2,000 combined baseline and count rows and 4 MB total JSON input. Output caps may require a smaller batch. The tool performs no external website, ERP or model API requests.

What does the report contain?

Field	Meaning
firstQuantity, recountQuantity	Exact normalized quantities, or null for a missing/ambiguous pass
rawRecountMinusFirst	Exact arithmetic on two unique quantities; not a cutoff certification
passConsistency	Observed equality/difference, or unresolved
varianceStatus	Comparable numeric difference, zero difference, or a specific unresolved reason
firstMinusBaseline, recountMinusBaseline	Exact deltas only when the cutoff is declared comparable and the relevant baseline/pass is unique
baselineRefs, firstRefs, recountRefs	All source row IDs and normalized source positions

The dataset contains one complete report with rows, findings, unused, scope and the full submitted rows. Summaries count keys rather than adding quantities across products or units. Other-session and undeclared-pass rows are retained as unused evidence. Related different-lot/unit hints show at most 20 references per key with a total count and an abbreviation flag; complete inputs remain in JSON.

CSV text is neutralized for spreadsheet formulas. Negative numeric delta strings consequently receive a leading apostrophe in CSV; JSON preserves the numeric text unchanged. HTML escapes supplied text and uses no external scripts. Download the HTML attachment and open it locally.

Pricing and recovery

The price is $0.25 per completed report, with platform usage included and no startup or per-dataset-item fee. A new run of the same input is a new paid report. An unknown or incomplete supplied-data result is still a useful completed report and is billable; invalid input or insufficient report budget produces no completed report event.

The Actor prepares bounded downloads before publishing one dataset item and one report event. Same-run resurrection verifies saved report integrity and input identity, repairs missing exports and avoids another completed-report charge. An interruption before dataset publication may leave partial downloads; resuming rewrites them before publication. Cloud publication and billing remain subject to Apify's platform behavior and the configured run budget.

Troubleshooting and support

If a key is ambiguous, keep its references and resolve the source export rather than selecting a convenient quantity. If your cutoff is unknown, supply better evidence or treat the result as unresolved. Use only authorized exports: raw inputs and reports are stored in your Apify run, so this is not a local-only privacy tool. See the API tab for programmatic access and the Issues tab for reproducible feedback using synthetic rows.