Elbuz imports products into the base catalog from a file, from your site via Elbuz Tunnel, or from a marketplace via API, then keeps the same cards updated without duplicates.
When you need this:
- Moving from another system — categories, brands, and photos come along.
- A supplier sends files regularly — a scheduled template updates existing cards, not copies.
- You set some prices yourself — “Fixed Price” on the card keeps import from changing them.
- A run broke off or imported the wrong thing — written data stays, a rerun doesn't duplicate.
Where to start: choose a method by what you have; make the first run on a small file.
What catalog import can do
- Four methods — ready-made file format templates, universal import, Elbuz Tunnel, and marketplace connections via .
- A template is set up once, then run manually or on a schedule.
- Repeat import recognizes a product by its identifier and updates the card instead of creating a copy.
- Card protection — “Do not update” and “Fixed Price” keep your edits.
- Deleting what's missing from the file — with a checkbox, the catalog mirrors the source.
- Run history shows what was added and updated and where errors occurred.
- Fixing a bad run — a source filter and bulk actions in the catalog.
Which import method to choose
Start from what you already have:
| What you have | What to choose |
|---|---|
| A live store on OpenCart, WooCommerce, PrestaShop, or another popular | Elbuz Tunnel — straight from the site's database, the most complete transfer |
| An export file in a known format (XML feed, 1C, a Prom spreadsheet) | A ready-made format template — YML, CommerceML, XLSX (EVO) |
| A file in any layout: your own spreadsheet, an export from another system | Universal import — you map the columns yourself |
| You already sell on a marketplace or keep records in an external system | Catalog import via API — with access keys, no files |
| You need orders from marketplaces, not products | Order import via API |
| Orders from your own OpenCart 3 site | The ready-made Elbuz API Sync module — orders arrive right after checkout |
Four ways to import data
1. From files with ready-made templates
The most common path: you have an export file from another system. Popular formats have ready-made templates with the columns already mapped:
- XML / YML (XML feed) — a universal XML product feed;
- CommerceML 2 (1C) — exchange with 1C: a products file and an offers file;
- XLSX (prom.ua, satu.kz) — an export spreadsheet from the EVO marketplaces;
- Elbuz Tunnel — a direct connection to the database of your existing store on OpenCart, WooCommerce, PrestaShop, and other CMSs, with no files at all.
2. Universal import
When there's no ready-made template for your file, use universal import. It reads any , , XLS, or file, and you specify which column or node of the file matches which catalog field.
It takes a bit more manual work at the start, but it fits any source — even a homemade export from someone else's accounting system. You map once: later files in the same format import with this template, no new mapping needed.
3. Via API — catalog
If you already sell on a marketplace or keep records in an external system, the catalog is pulled directly through the API — no files, using access keys. This is how Rozetka, Shopify, WooCommerce, Amazon, eBay, Allegro, eMAG, Kaufland, OLX, Etsy, TikTok Shop, Wix, as well as Horoshop and the SalesDrive connect.
You get the keys in the marketplace or system account — usually in the developer or integrations section — and paste them into the template form. Elbuz then pulls products, categories, and images just as it does from a file. Each integration has its own form and keys — the articles are collected in the catalog import via API overview.
4. Via API — orders only
Some integrations import orders rather than products — from marketplaces, POS systems, and banks: Kaspi.kz, EVO/Prom (prom.ua, satu.kz), Epicentr, Forte Market, ceneo.pl, the dotykacka POS, PrivatBank statements. They don't fill the catalog — see the order import via API overview.
Catalog or orders
Some sources bring a catalog, others bring orders, and some APIs can do both. In the template you turn on what to pull: products, orders, or both. File templates and universal import are always about the catalog; some APIs are about orders only.
How import works: from template to catalog
The steps are the same for any source:
- Create a template and choose the source type.
- For a file, specify the file or a link and check the column mapping; for an API, enter the access keys.
- Turn on what to import: products, categories, manufacturers.
- Run the template. The first run creates products, later runs update them.
- Once everything works, put the template on a schedule.
- Action or result
- Decision: the path depends on the answer
One template, one source. Each supplier or marketplace gets its own template with its own schedule: one updates every hour, another once a day. But they share data: if two templates bring the same SKU, they write to one card (two templates for the same product).
What ends up in the catalog
Depending on the source and template, the catalog receives categories (the section tree), manufacturers (brands), images, specifications and attributes, prices, and stock. What to take is set differently:
- in Elbuz Tunnel and marketplace connections, photos, attributes, options, manufacturers, reviews, and orders are turned on with separate checkboxes in the template form;
- in file format templates, the set is fixed;
- the most precise control — down to a single field — is in universal import, through column mapping.
What happens to fields you edited by hand is covered in what a repeat run overwrites.
Service categories and brands that import creates
Templates handle a product that has no category in the source differently:
- Universal import and marketplace connections don't lose such a product: they put it in a service category with the source code in its name. It's created disabled and with export off — its contents won't reach the storefront until you sort the products into your own sections.
- File format templates (XML/YML, CommerceML, XLSX) don't add a product without a recognized category — silently, with no line in the log. This is the most common cause of “the file has more items than were added to the catalog”.
You can sort them in bulk: select the products, then “Additional Operations” → “Change product category” → choose the main category (there can be only one main category). After that the service category is empty and can be deleted.
Don't delete a category while it still contains products. Deleting a category takes with it all products for which it's the main category — along with descriptions, photos, and attributes. The program shows “Please wait, deleting the category and its products...” but doesn't ask separately. Move the products first, then delete.
Deleting a brand doesn't touch the products — their manufacturer field becomes empty. Markup rules tied to that brand are deleted along with it. If the brand still comes in the file, the next run will create it again.
When imported data moves on
At the end of a run the program finalizes the data itself — log lines “Updating values based on formulas”, “Data integrity check”, “Product indexing”. Then the product appears in the base catalog list. Smart search is only queued for reindexing as a separate task, so search shows a new product with a delay.
Import sends nothing to your site or marketplaces. That's done by export — separate templates with their own schedule. Until an export runs, the storefront shows the old data. Products didn't appear on a marketplace after import? Check the export log.
How to run a template and check that the import went through
For a one-off run, use the “Import” button on the template card. For automatic runs, use the “Schedule setup” button on the template card; the template must be “Active”. On a schedule the program fetches data by link or keys, so a file from your computer only works for manual runs. Universal import has no schedule: you run it manually.
Every run leaves a trace. The “Run history” button is in the template form and on the toolbar above the template list when a row is selected. It shows when the import ran, how many products or orders were processed, added, and updated, and any errors. Problem rows are color-coded, with the source's reason text next to them.
For the first few runs, open the log every time, even when everything seems fine. It immediately catches a changed password or token in the external system, a source that returned fewer products than expected, and an empty or outdated file.
Limits of a single run
There's no cap on the number of products. The file isn't loaded into memory whole but read sequentially. Universal import writes to the database in batches of 2,000 rows; file format templates first fill staging tables and move everything into the catalog in one go. Catalog size is limited by time, not a cap.
A run of the EVO marketplace XLSX export and of universal import gets two hours. A universal import that runs out of time leaves the catalog partially imported — the written batches are already there. XLSX is first read into staging tables, so a run stopped by the limit while reading doesn't change the catalog.
“Run history” shows how long your own runs take: for the last 30 days, the number of runs, “Average duration”, “Longest run”, and how many runs were “With errors”. After a week you'll have your own numbers for choosing a schedule.
The window isn't locked during import: the run log lives in the “Operations” panel, and you can close the template card. Scheduled runs happen on the server and don't need a browser.
Above the template list is a summary bar: “Templates”, “Active”, “Running”, “Last import”, and “Not imported recently”. The last one counts active templates that haven't run for more than 7 days — so you can see that a source has gone quiet.
Disable a template or delete it
At the top of the template card there's a status badge — “Active” / “Disabled”; the “State” column in the template list toggles the same thing.
A disabled template skips its schedule and isn't picked up by “Import all templates”. Settings, keys, mapping, schedule, and already imported products stay. Turn it back on and the schedule continues. The “Import” button on the card works for a disabled template too: disabling stops only automatic runs.
Deleting a template also removes its schedule, and for Elbuz Tunnel the site database credentials, filtering rules, and matching too. Imported products, categories, and manufacturers stay in the catalog.
A deleted template can't be restored through the interface — you'll have to set it up again. To pause a template, disable it rather than deleting it.
Repeat import: updates without duplicates
Products don't get duplicated on each import. Elbuz recognizes an item by its identifier — SKU, code, or external ID from the source — and updates it. So you can run a template daily: existing items update, new ones are added. If the source's identifiers change between exports, the program takes old products for new ones and creates duplicates.
What a repeat run overwrites
The main question before a second import is whether it will wipe out what you edited by hand. The answer depends on the field:
| Fields | Behavior on a repeat run |
|---|---|
| Availability and quantity, price and old price, name and description | Always taken from the source. No value in the file — the card gets an empty value: a product without a price gets zero, without availability — the default status (the first one marked as “In stock”) and its quantity |
| SKU, model, , ISBN, warranty, tags, category, manufacturer, product link in the source | In the XML/YML template — only if the file value isn't empty. In CommerceML and XLSX (EVO) they're written as is: an empty value overwrites yours |
| Photos, attributes, options, manufacturers, reviews, orders | In Elbuz Tunnel and marketplace connections — by the template checkboxes. In file format templates, photos, attributes, and manufacturers are always taken |
| Custom fields you created yourself | Updated only with the “Update custom fields” checkbox (available in the XML/YML template) |
So an incomplete file — names only or photos only — won't do for a repeat run. The exception is CommerceML exchange with a single file: an offers file (offers.xml) without a products file updates only price, stock, and availability status, and a products file (import.xml) without offers updates only catalog fields.
How to lock your own value
Where: product card → “Statuses and flags” section → “Import protection” group.
- “Do not update” — import no longer changes this card: not the price, availability, SKU, category, manufacturer, or name and description. This applies to file format templates, Elbuz Tunnel, universal import, and API.
- “Fixed Price” — no import touches the price or old price: file format templates (XML/YML, CommerceML, XLSX), Elbuz Tunnel, universal import, and API. The other card fields update as usual.
1
Do not update2
Fixed Price- Do not update — the import skips this product entirely.
- Fixed Price — the import leaves the product's price as it is.
Templates handle photos on existing cards differently:
- universal import and API add images only to new products — an existing card keeps your photos;
- the XML/YML and CommerceML templates add images the card didn't have yet and leave the old ones alone;
- the XLSX template for the EVO marketplaces replaces photos entirely: only those in the export remain.
You can update only prices and availability, leaving everything else alone, with universal import mapping — see the scenario below.
Two templates for the same product
The program recognizes a product by the identifier from the source, not by the template that created it. If two templates bring the same SKU, they work with one card, and the value from whichever ran last stays. There's no “source priority” setting.
To separate them, use the “Check uniqueness of records based on this import source” checkbox on the template card. With it, a run looks for an existing product only among its own, so identical identifiers from different sources become separate cards. It's in the XML/YML and Elbuz Tunnel templates; CommerceML, XLSX (EVO), and API connection forms lack it.
In XML/YML templates, a field that's empty in the second source doesn't overwrite the first source's value. This doesn't apply to price, availability, name, and description: two full files for one assortment in one card keep overwriting each other.
A product disappeared from the source
A run compares the source with the catalog in one direction only: what's in the file is updated or added. A product no longer in the file stays as it was by default — with its price, availability, and photos. The card isn't disabled and doesn't disappear on its own.
The “Delete entries in the base catalog that are not in the imported file” checkbox changes that (Elbuz Tunnel: “...that do not exist on the site”). With it, a run deletes everything missing from the file, plus descriptions, photos, attributes, category links, and product questions. Their price list matches are lost: the row is left without a product.
The checkbox works on the whole base catalog, not just this template's products. “Check uniqueness of records based on this import source” narrows the scope: other templates' products stay. Cards created by hand, from a price list, or by the parser are subject to deletion either way. Deleted items can't be recovered.
The XML/YML, XLSX (EVO), and CommerceML templates protect against an empty file: if the file contains no products at all, no products are deleted. CommerceML does the same with categories: without a classifier in the file (exchange with an offers file only), no category is deleted. The Elbuz Tunnel template has no such check.
- Action or result
- Decision: the path depends on the answer
Among API connections, only the Horoshop and SalesDrive forms have a checkbox for deleting missing items:
- SalesDrive — deletes: its export (a YML export by the link in the template) is imported the same way as an XML/YML file. There's no “Check uniqueness of records…” checkbox here — the whole base catalog is subject to deletion. Without a link to the YML export, SalesDrive pulls only orders and deletes nothing.
- Horoshop — the checkbox deletes nothing.
- Other marketplaces don't have the checkbox: a product taken off sale stays in the catalog until you remove it yourself.
If a run broke off or was stopped
A run writes to the catalog as it goes, not on an “all or nothing” basis. If the connection dropped, time ran out, or someone clicked “Stop”, what was written stays written, and what wasn't reached isn't done.
Rerunning after a break is safe: products already added will be found by identifier and updated, the rest will be added. This doesn't create duplicates.
The “Stop” button in the “Operations” panel sets a flag, and work stops at the nearest checkpoint. The log shows “Operation aborted by user!”, and the operation in the “Operations” panel turns yellow. Formulas, the integrity check, and indexing don't run after a stop, there will be no “Data import completed!” line, and the template's last import date won't change.
If deleting items missing from the file is turned on in the template, the deletion happens before new data is written. A run broken off right after this stage leaves the catalog already without the deleted products but not yet updated.
- Action or result
- Decision: the path depends on the answer
Don't start two runs at once: file templates share a staging area, so the second run overwrites the first one's data. The schedule checks this itself — a busy moment gives “Scheduled import canceled due to another template being imported (data processing conflict)” in the log, and the task waits for its next time. Manual runs skip this check.
A scheduled run that died silently doesn't hold the lock: after 15 minutes of silence the task is released, and next time the schedule runs as usual.
If you imported the wrong thing
First, assess the scope. The base catalog list has columns for this:
- “Source: name” and “Source: ID” — which template created the card. Filtering by this column gives the full list of products created by the bad run.
- “Date updated” — when the card was last changed: by an import, bulk actions, or saving the card. It also catches cards the run didn't create but only updated: their source stays the old one.
Next, bulk actions from the “Additional Operations” menu (the three-dot button in a row, or right-click the list):
- “Delete product” — for anything extra; for selected rows it asks “Delete the selected products?”;
- “Change product category” — move products to the right section;
- “Set value in bulk” and “Find and replace” — fix or clear a specific field.
1
Delete product2
Change product category3
Set value in bulk- Delete product — removes the product (it goes to the trash).
- Change product category — moves the selected products to another category.
- Set value in bulk — writes one value into a field for many products at once.
A repeat run of the correct file restores overwritten prices, availability, names, and descriptions — these fields are always taken from the source. Fields the bad file filled and the correct one leaves empty (SKU, model, warranty, tags) won't clear themselves in the XML/YML template: clear them with “Set value in bulk” and an empty value.
Deleted items can't be recovered. If the run had the delete-missing checkbox on, the products are gone from the database for good. A repeat run will create them as new cards: with new internal identifiers, without manual edits, and without matches to price lists.
Scenarios
First import: start with a small file
What to do. Choose a method from the table. Set up a template on a small file or a single category, run it, and check the cards and the run history.
Result. A mapping or category error shows up on a dozen products, not the whole catalog. Once the cards are right, import everything and put the template on a schedule.
A feed drives the assortment: what's removed from the feed must disappear from the catalog
What to turn on. In the XML/YML template — the “Delete entries in the base catalog that are not in the imported file” and “Check uniqueness of records based on this import source” checkboxes.
Result. Each run deletes from the base catalog the products not in the file, along with descriptions, photos, and attributes. Other templates' products stay. An empty file won't wipe the catalog: the deletion stage is skipped.
Keep in mind. The second checkbox doesn't protect cards created by hand, from a price list, or by the parser. If the first checkbox is grayed out and can't be clicked, your group doesn't have the right to delete products during import.
Two suppliers with the same SKUs, and you need separate cards
What to turn on. In each supplier's template (XML/YML or Elbuz Tunnel) — the “Check uniqueness of records based on this import source” checkbox.
Result. A run looks for an existing card only among those created by this template, plus those created by hand, from a price list, or by the parser. The same identifier from two suppliers gives two cards.
Keep in mind. A card created by hand with the same identifier will be updated by both templates — it counts as “their own” for each.
Update only prices and availability, leaving descriptions, photos, and other fields alone
What to turn on. Universal import → “Field mapping” tab → “Product fields” block: keep price and quantity in the mapping, plus the manufacturer SKU, name, and product link in the source.
Result. Only the mapped fields change on existing cards. An empty price or SKU cell in the file doesn't overwrite the value, but an empty quantity does: zero quantity means “Out of stock”.
Keep in mind. The program finds the existing card by the combination of manufacturer SKU, name, and link. Remove them from the mapping and every run will create new cards.
A supplier is on pause: stop automatic import without losing the settings
What to turn on. On the template card, switch the status badge to “Disabled” (or do the same in the “State” column of the template list).
Result. The template skips its schedule and the “Import all templates” command. Settings, schedule, and products stay. Switch back to “Active” and the template runs on schedule again. More: disable a template or delete it.
What to do when import doesn't behave as expected
| What you see | Why | What to do |
|---|---|---|
| The file has more items than were added to the catalog | File format templates don't add a product without a recognized category | Check the categories in the file (service categories) |
| “Scheduled import canceled due to another template being imported (data processing conflict)” | Another run was in progress at that moment | Space out the template schedules (a broken-off run) |
| “Operation aborted by user!” | The run was stopped with the “Stop” button | Run the template again — there won't be duplicates (a broken-off run) |
| Products were duplicated after a repeat run | Identifiers in the source changed, or two templates were separated with the uniqueness checkbox | Find the copies with the “Source: name” filter (imported the wrong thing) |
| A manually set price changed after a run | Price is always taken from the source | Turn on “Fixed Price” on the card (lock your own value) |
| The “Delete entries…” checkbox is grayed out and can't be clicked | The group doesn't have the right to delete during import | Give the group the right in “Group permissions” (who has access) |
| The catalog has new data, but the site still shows the old | Import doesn't export to the site | Run an export template (when imported data moves on) |
FAQ
How is this different from importing a supplier's price list?
A supplier's price list is imported for price comparison and dropshipping — that's the “Suppliers” section. Here you fill your own base catalog: with your products from a file, a marketplace, or an accounting system.
Can I import the catalog from my marketplace store without a file?
Yes, if the marketplace has an API integration (Rozetka, Shopify, WooCommerce, and others). Elbuz pulls products using access keys, without files — catalog import via API.
I connected an API, but orders arrive instead of products — why?
Some integrations bring only orders (POS systems, banks, some marketplaces). In integrations that handle both catalog and orders, check the template checkboxes: product import must be turned on.
Do I have to import manually every time?
No. A tested template with a link or keys is put on a schedule, and the import runs on its own. The exception is universal import: you run it manually (running and scheduling).
I edited a price by hand — will a repeat run wipe it out?
Yes, price and availability are taken from the source every time. To keep your value, turn on “Fixed Price” or “Do not update” on the card (how to lock your own value).
Two sources send the same IDs and overwrite each other. Can I import them as different products?
Yes: in each source's template, turn on “Check uniqueness of records based on this import source”. The checkbox exists in the XML/YML and Elbuz Tunnel templates (two templates for the same product).
Can I connect a Google Sheet directly as a catalog import source?
No. Google Sheets are read directly by price lists: a source with the “Link” method (price list import), followed by matching with the catalog. The other way is to export the sheet to a file and import it with a catalog template.
Where it is and who has access
- Menu
- Directories → Operations → Data import
- Access right
- the “Import” window in “Group permissions”
- Delete checkbox
- a separate right, “Flag: delete categories and products in the base catalog that are not in the uploaded file”
- Have ready
- a file or a link to it, access to your site, or the marketplace's API keys
The “Import categories, products, manufacturers to the base catalog” window works through templates. A template is the saved setup of one import: where to get the data, in what format, by what rules, and what exactly to pull.
You set a template up once, then run it manually or put it on a schedule. Each rerun updates the catalog by the same rules.
1
Template name2
Source type- Template name — each row is one saved way of loading data into the base catalog.
- Source type — the line under the name shows how the data arrives (XML, Elbuz Tunnel, universal file import); one list holds them all.
Related topics
- If you have neither a file nor a marketplace and need to collect products from website pages — site parser.
- If the catalog is filled and it's time to connect suppliers — Price lists: where to start and where to go next.
- If your file format is non-standard — universal import with manual mapping.
- If you sell on a marketplace and want to pull products without files — catalog import via API.
