Running a programme · For a brand's operations team

Uploading batches from a spreadsheet

When a brand orders twenty-five batches at once, typing them into a form is where mistakes happen. The upload path takes a spreadsheet in three steps, and each step exists to stop a specific mistake.

Three steps #

The three steps of a batch upload
Step	What it does	What it stops
Template	An Excel file downloaded from the product's own Bulk upload tab. Its `product` dropdown holds that one product, and `batch_kind` is a dropdown too.	A typed product name that matches nothing — and a batch generated for the wrong product.
Preview	Parses the file, validates every row, resolves the products and prices the file against the credit balance. It writes nothing.	A surprise.
Commit	Asks you to retype the label total, the way a bank asks you to retype an amount.	A mis-click.

Template
What it does
An Excel file downloaded from the product's own Bulk upload tab. Its product dropdown holds that one product, and batch_kind is a dropdown too.
What it stops
A typed product name that matches nothing — and a batch generated for the wrong product.
Preview
What it does
Parses the file, validates every row, resolves the products and prices the file against the credit balance. It writes nothing.
What it stops
A surprise.
Commit
What it does
Asks you to retype the label total, the way a bank asks you to retype an amount.
What it stops
A mis-click.

After the commit the background worker writes the batches — always the worker, never the request, because twenty-five rows of 10,000 labels is a quarter of a million labels.

The columns #

product, quantity, batch_kind, programme, lot_number, batch_name, scan_limit, hidden_code_format, mfg_date, expiry_date, offer_end_date, discount_price_rupees, fact1_name, fact1_value, fact2_name, fact2_value

Only product and quantity are required. product takes the product's id or its exact name (capitals do not matter) — and it has to be the product whose page you uploaded from: a row for any other product is rejected by name. Only Excel (.xlsx) is accepted; a CSV is refused with a sentence saying so.

Limits and formats #

A file holds up to 1,000 rows and 2 MB.
Dates are YYYY-MM-DD and nothing else. A date means the end of that day, IST.
discount_price_rupees is a whole number of rupees. The verdict page shows it as Special price.
batch_kind takes a value from the template's dropdown. The values are not listed on this page on purpose: they are the console's own list, and the template always carries the current one.
The expiry date, the offer end date and the special price take three of the verdict page's five fact slots, so a row has two left for facts you name. A fact3_name column you add is read, but a sixth fact is rejected with that reason.
The dropdowns are typing help, not a gate: a pasted value skips them. Every row is checked again on upload.

When a row is rejected #

The preview names the row and the reason. Fix the file and preview again; nothing was written, so there is nothing to undo.

The preview also prices the file against your balance and shows any shortfall in red before you commit. If you commit a file the balance cannot cover, the rows are generated one by one until the credits run out, and the report lists the rest as failed. A row is never half-generated, and the balance never goes below zero.

Uploading again #

The same file uploaded twice generates nothing twice, and spends nothing the second time.
Change one row and upload the file again, and only that row is generated again, as a new batch. The batch the old row made stays as it was.
Two identical rows in one file are rejected, because they would be generated once. Give each its own lot number.

Last reviewed 19 Oct 2026 · Something wrong? connect@avrascan.com