QuickBooks Historical Import uses the Arrow Down To Line icon and is separate from QuickBooks Online OAuth/synchronization. It is a direct-write migration tool with no row preview, dry run, duplicate detection, undo, purge, or rollback. Contrary to the screen description, the current function reports products_created 0 and customers_created 0: it creates completed sales, denormalized sale lines, payments, and missing payment-type labels; it only links existing products/customers when exact matching succeeds. Re-importing the same receipt can create a duplicate QB-<receipt> sale.

Harbor & Hearth QuickBooks Historical Import screen showing instructions, pristine file selection, and import controls in CentriPOS
Harbor & Hearth live CentriPOS demo
Before you beginUse an authorized administrator account and confirm the business shown in the top bar. Record the previous value before changing a live operational setting.

Step by step

  1. Go to Admin → QuickBooks Historical Import in the intended business. Confirm the Arrow Down To Line navigation icon and Upload QB Export card. Do not confuse this with Admin → QuickBooks Online or the Lightspeed migration console.
  2. Prepare a reviewed export in the importer’s exact positional structure, not an arbitrary QuickBooks report. It reads only the first worksheet and skips the first row. A sale header must have Receipt # in column C and the literal Sales in column F; it reads Date A, customer G, total L, payment M, and store O. Following item rows read SKU D, product name E, quantity I, and line amount K.
  3. The file control accepts .xlsx, .xls, and .csv even though its visible prompt says .csv or .xlsx. Parsing loads SheetJS from its CDN in a browser worker, ignores formulas/styles/date typing, reads at most 200,000 worksheet rows, and uses only the first sheet. Confirm network/privacy policy permits this client-side parser dependency.
  4. Create a small controlled file externally containing a few representative receipts in the exact layout. The page has no date range, row selector, receipt preview, column mapping, header validation, company identifier, duplicate scan, or file-content redaction before import.
  5. Choose the file and wait for Reading, Parsing spreadsheet, Converting to rows, and Parsing sales to finish. The screen shows only parsed sale and line-item counts, not receipt/customer/item/payment details. “No sales found” means no qualifying headers with item rows were detected; do not import by changing random columns.
  6. Be aware that the UI’s cap toast says 100,000 rows while the current worker actually sets and detects a 200,000-row cap. Treat any capped result as incomplete and split/reconcile the source deliberately.
  7. Review Map QB store codes to locations. The importer detects nonblank store strings from parsed sales and may auto-map by case-insensitive substring against location names. Auto-match is a suggestion, not proof. Select the exact destination for every code; Import remains disabled unless at least one store was detected and all detected codes are mapped.
  8. Before Import, audit existing sales for each planned QB-<receipt number>. The importer does not check for an existing sale number/source receipt and the database write path does not upsert. Running the same file twice can create duplicate completed sales, lines, and payments.
  9. Understand customer matching. It normalizes the export name and existing customer/company names by lowercasing, changing & to and, removing punctuation, and collapsing spaces. Exactly one matched customer ID is attached. Zero matches leaves the sale customer null and records the raw name in the sale note; multiple matches also leave it null and report Ambiguous. No customer is created.
  10. Understand product matching. Each item first looks for an exact tenant product SKU, then an exact product name. A match links product_id. No match still creates a receipt line whose name is suffixed (QB Import), with the source SKU, but creates no catalog product. Variant matching is not performed and cost is always 0.
  11. Understand payment behavior. The importer matches payment types case-insensitively by name and creates an enabled payment_type row when none exists. It then creates one completed payment for the sale total or computed subtotal. Review unsupported/legacy tender names before import so they do not silently become new enabled reporting methods.
  12. Understand historical sale values. Each record is inserted Completed with sale number QB-<receipt>, destination location, subtotal from line amounts, discount and tax 0, total/amount paid from nonzero export total or subtotal, change and balance 0, and a QB Import note. Date text is parsed as M/D/YY or M/D/YYYY at local noon then stored as an ISO timestamp. No register, register session, tax reconstruction, discounts, returns, account balance, or inventory movement is derived here.
  13. Choose Import <N> sales once. The browser sends sequential batches of 50 to the function and updates a percentage. There is no cancel button. A failed batch or individual receipt does not roll back earlier successful batches/rows, and a sale can exist even if one of its later line/payment writes reports a problem.
  14. Read Import Complete as counts, not reconciliation: sales, line items, products created (currently 0), customers matched, unmatched, ambiguous, and up to 20 displayed warnings. The client aggregates batch warnings and appends up to 25 ambiguous/unmatched names; those names are customer data and must not be copied publicly.
  15. If any warning/error occurs, stop. Check whether the affected QB sale and partial child rows already exist before retrying. There is no Import History, job ID, undo, delete-import, or source-batch identifier on this screen beyond the QB sale number/note. Use an approved surgical data-repair process if rollback is required.
  16. After a clean controlled batch, reconcile source versus CentriPOS receipt count/IDs, location, customer link/unmatched note, item names/SKUs/quantities/amounts, subtotal/total, payment type/amount, timestamps, zero tax/discount/cost assumptions, and reporting behavior. Only then proceed with another nonoverlapping reviewed file.
  17. Choose Import Another File only after preserving evidence of the completed batch. It clears the page state; it does not undo or deduplicate anything already written. Maintain an external import ledger of company, file hash/name, row/receipt range, store mappings, time, operator, counts, warnings, and reconciliation signoff.

Screen and contract reference

Control or areaWhat it means
Arrow Down To Line iconExact CentriPOS navigation icon for QuickBooks Historical Import.
Upload QB ExportDirect-write historical receipt importer, separate from QBO sync.
Accepted file control.xlsx, .xls, or .csv; first worksheet only.
Expected header rowDate A, Receipt # C, Sales marker F, customer G, total L, payment M, store O.
Expected item rowSKU D, product name E, quantity I, line amount K beneath its header.
200,000-row parser capActual worker limit; the UI toast currently misstates 100,000.
Parsed countsSale/header groups with at least one product-name line; not a record preview.
Store mappingEvery detected source store must map to one current CentriPOS location; substring auto-map requires review.
Import buttonSends direct-write batches of 50; no dry run, confirmation, cancel, or rollback.
QB-<receipt>Generated CentriPOS sale number; no existing-sale duplicate check is performed.
Customer matchingExact one-result normalized name/company match only; unmatched/ambiguous sales have no customer link.
No customer creationCurrent function returns customers_created 0.
Product matchingExact SKU first, then exact name; no variant match.
No product creationUnmatched items become denormalized sale lines suffixed (QB Import); products_created remains 0.
Payment type creationUnknown nonblank method creates an enabled tenant payment-type label.
Imported sale assumptionsCompleted, zero tax/discount/cost/change/balance, fully paid, no register/session or inventory reconstruction.
Batch partial-write riskEarlier batches and parent/child writes are not rolled back when a later operation fails.
Import CompleteAggregated counts and warnings; not proof of source reconciliation.
Import Another FileClears local UI state only; previously imported data remains.
No built-in history/undoScreen provides no import job ledger, batch ID, dedupe, purge, or rollback.
Harbor & Hearth screenshotReal Amelia Grant capture of /admin/qb-import in the pristine upload state. It shows no selected file and exposes no accounting, customer, product, receipt, or payment data.

Expected result

The reviewed historical import creates only the intended customer-safe accounting history, records source references and exceptions, and does not duplicate existing CentriPOS records.

Troubleshooting

  • Confirm you are changing the intended business and location before saving.
  • Record the previous setting when a change affects live selling, payments, tax, or integrations.
  • Use a controlled test record after saving before rolling the setting out broadly.

More guides

Build an itemized payment in MoneyUse Products or Amount & description to build one receipt, attach the customer, and follow the appropriate payment or service-estimate workflow.
Take a one-time payment in Virtual TerminalReview a one-time amount, tax and contact at the correct payment station, use its verified secure checkout, and recover the original receipt after an uncertain result.
Review booking payments, refunds and earned careKeep a booking’s customer payments, unused advances, completed care, cancellation fees and refunds distinct, then read a correctly scoped accrual report.
Configure booking properties and customer pagesSet accommodations, service packages, availability, required intake and payment terms, then review the correctly branded booking page before sharing it.