Supplier Catalog Change Review

Compare a previous supplier catalog with an incoming CSV and get a traceable report of additions, absences, price changes, stock changes, and other field changes. Invalid data is explained before it can produce a misleading comparison.

This initial release supports two explicit CSV formats. It is useful as a bounded step in a catalog-processing workflow. It does not connect to a supplier account or update a store.

What you receive

• Report JSON: authoritative validation status, counts, record changes, source evidence, input hashes, and custom-event billing status.
• Changes CSV: a review table of the records that differ.
• Exceptions CSV: input problems and reasons a comparison was blocked.

All three files appear in the run's Output tab. Always inspect the JSON status: an empty changes file can mean no changes, validation-only mode, or a blocked comparison. It does not by itself establish that two catalogs match.

A successful validation-only run also writes one summary to the run's default dataset: the profile, parsed and valid row counts, issue count, comparison status, and keys for the detailed report files. This summary does not request a custom comparison event. Comparison runs continue to deliver their results in the three files above.

Input

Choose a supported profile, paste the current CSV contents, and optionally paste the previous CSV contents. API callers send the same values as JSON strings. No file URL, account credential, local path, or custom executable configuration is needed or accepted.

For comparison, confirm that both files are complete snapshots of the same catalog scope. The tool cannot independently discover whether an export omitted otherwise valid rows or verify which snapshot was captured earlier.

Field	Meaning
profileId	generic_catalog (default) or synthetic_building_materials
currentCsv	Required inline CSV text, including its header
baselineCsv	Optional previous full CSV; omit the field for validation only
snapshotScopeConfirmed	Must be true when a baseline is supplied

Generic format

sku,name,price,stock,currency
001,Example item,10.00,25,USD
002,Another item,12.50,8,USD

Every listed column and row value is required. SKU is exact, case-sensitive text; leading zeros are preserved. Price is a nonnegative decimal with at most two places; stock is a nonnegative integer; currency must be USD.

Synthetic building-materials format

Supplier Code,Item Code,Description,Sell Unit,Unit Price,Available Units,Currency
SUP-A,ITEM-001,Example material,each,10.00,25,USD

The exact supplier code plus item code identify a record. Both are required. The optional extra column Internal Note is deliberately ignored. The sell-unit field provides price context: a unit change prevents an incomparable price delta. This profile is an example schema, not verified compatibility with a named vendor.

Interpretation and limits

Identifiers are never guessed, trimmed, case-folded, or fuzzily matched. Duplicate identifiers, invalid values, malformed CSV, unexpected columns, or incomplete reads block the entire comparison. The output explains the problem and returns null comparison counts rather than misleading zeros.

An item marked removed is absent from the supplied valid snapshot. It is not a destination deletion instruction or proof that a supplier discontinued it. Changed identifiers appear as an addition and an absence.

Each snapshot is limited to 2,000,000 UTF-8 bytes and 10,000 data rows. Headers must match exactly. The profiles reject empty snapshots, price rounding, thousands separators, currency symbols, exponent notation, negative stock, and unknown currency codes. No currency or unit conversion is performed.

CSV review files neutralize formula-leading cells. JSON retains the original relevant values and represents decimal amounts as strings to preserve precision.

Pricing behavior

The proposed sale unit is one completed, valid comparison. The actual price is the event price displayed by Apify. The custom event is requested only after the report and review files have been stored.

Validation-only reports and invalid inputs do not trigger the custom comparison event. Total platform charges depend on the published pricing configuration. The intended release configuration contains only catalog-comparison, with no automatic startup or dataset charge and no passed-through usage charge.

A spending limit must cover one event. A fresh run is a new execution, even if it uses identical input. Within a resumed run, recorded completed work is reused. If a charge response is ambiguous, automatic charging stops for review.

Data handling

The tool processes the CSV in an isolated temporary directory and removes those temporary files when processing ends. Inputs and result artifacts also exist in Apify run storage and follow that account's retention and deletion settings. Delete the run's input and output storage when it is no longer needed. This package does not configure an account-wide retention policy or automatically delete stored run results. Routine application status messages contain summaries, not CSV contents.

Scope

This tool produces review information. It does not submit imports, change prices or inventory, verify external product identity, or determine why stock changed. Use a new run for each new pair of snapshots.