↑↓ selectEnter openEsc close

    1. Elbuz
    2. Manual matching

    Matched products directory: how to check and break a link to a price list

    25 min readUpdated

    Elbuz's directory stores every confirmed “price list row — catalog product” pair and restores it on each import, so each item is confirmed once. Check and break pairs here.

    When you need this:

    • A card got someone else's price — break the pair with the trash icon in its row.
    • A product temporarily left the price list — the pair stays as a lost link and returns with it.
    • A supplier was bulk-matched wrong — “Delete” removes all its pairs.
    • Before a risky change — export pairs to a file to import back later.

    Where to start: source → arrow next to “Product Matching” → “Matched products”.

    What the directory does

    You don't create new pairs here: that happens in the product matching window, through AI matching or from a file.

    How many records the buttons affect: one price list, a supplier or the whole directory

    The window can show one price list or all price lists at once. It takes the mode from the address on opening: a source header passes the price list number, the “All products” tab passes zero. There's no mode switch, the heading doesn't mention the mode, and the window name is the same in both.

    You can tell the modes apart by the “Show lost links” toggle on the right side of the toolbar. It appears only in single price list mode. No toggle means you're in all price lists mode. A second, less reliable sign: the “Supplier” column shows different suppliers, not just one.

    Besides the price list number, the window also takes the supplier number from the address, and together they decide how many records the buttons affect. That gives three scopes, one for each way of opening the window.

    Three scopes the buttons work in
    How the directory was openedScope of operations
    From the source headerrecords of this price list
    From the “All products” tab of the selected supplierrecords of this supplier — across all of its price lists
    From the “All products” tab of the top node “All sources”the shop's entire directory — all suppliers, all price lists
    Button bar of the Matched products window
    1Delete
    2Start product matching (wand)
    3Save found products as manual (disk)
    4Export to XLSX (arrow)
    5Import from CSV (arrow)
    6Show lost links toggle
    1. Delete clears the whole directory of this price list.
    2. The wand restarts product matching.
    3. The disk saves automatically found products as manually matched.
    4. The first arrow exports the matches to an XLSX file.
    5. The second arrow imports manual matches from a CSV file.
    6. The toggle shows matches whose price list row is gone; it is off here.
    The same buttons in two modes
    What you doSingle price list modeAll price lists mode
    “Show lost links” togglepresentnot shown; lost links are always displayed
    “Delete” buttonclears records of this price listclears records of the selected supplier — across all of its price lists; opened from “All sources” — the whole directory
    “Start product matching”rebuilds the links of one price listgoes through all active price lists of active suppliers
    “Save automatically found products as manually matched”for one price listfor all price lists at once
    Export to XLSXrows of this price listthe supplier's entire directory, or from “All sources” — the whole directory (what goes into the file)
    Filtering by suppliernarrows the list to the selected supplier's records; directory records without a supplier stay in itthe same: the list shows the selected supplier's records, both manual and automatic, plus directory records without a supplier

    The confirmation dialog names the cleanup scope — read it before you confirm. The question is always the same: “Delete all product match records?”.

    Without a price list number, a red scope line appears below: “All links of the selected supplier will be deleted — across all of its price lists”, or, if opened from “All sources”, a warning about all links in the directory — for every supplier and every price list of the shop. No red line means one price list.

    No dialog shows a record count. There's no undo either: nothing to restore deleted records from. The work log shows that it happened — a “Matching directory cleared:” line with the scope name.

    In all price lists mode, opening the window again doesn't change the scope. For each price list Elbuz opens a separate window: the directory of price list A and the directory of price list B can be open side by side.

    But there's only one all price lists window. You opened the directory from one supplier's “All products” tab, then click “Matched products” for another supplier or in the top node “All sources” — Elbuz doesn't load a new window, it brings the open one to the front, still with the first supplier's data.

    Neither the heading nor the list changes, and there's no message. To switch to another supplier or to the whole directory, close the window first.

    How to read the list of pairs

    Each row in the list is one pair. A row has twenty-four columns, and half of them are rarely needed, so below are only the ones people come here for.

    • Trash icon in the first column — breaks this link.
    • “Match type” — how the pair was found: manually, by manufacturer SKU, by product name, by barcode, or by other matching rules.
    • “Link to price list lost” — a checkbox meaning there's no live price list row for this pair anymore.
    • “Matched via AI” — a checkbox, the only way to tell a pair picked by the model from one confirmed by hand.
    • “1 price list product matched to several catalog products” — the group link checkbox: one price list row passes its price and stock to several catalog products.
    • “Price”, “Stock status (from price list)” and “Category” — come only from a live price list row. They're empty for lost links.
    • “Date added” and “Username” — when the pair was created and by whom. The name is shown as “Name (login)”.
    The Matched products list window with the blue help banner, the button toolbar and the first rows of the grid
    1Help banner
    2Match type column
    3Link to price list lost column
    4Trash icon in the first column
    1. The blue banner explains what the directory is and what the match type means.
    2. The Match type column shows how the pair was found: manually, by SKU or by product name.
    3. The Link to price list lost column is checked when the price list row no longer exists.
    4. The trash icon in the first column breaks the link for that row.

    Four more columns — “Catalog: Manufacturer SKU”, “Catalog: Name”, “Price: Manufacturer SKU” and “Price: Name” — are an archive. Their values were written to the directory when the pair was created and never change. They show what the product was called back then, which is the only way to understand why these two items were linked.

    What this list doesn't have: pages (scrolling is continuous), column settings and saved views, in-cell editing, or selecting multiple rows with checkboxes. There's exactly one bulk action — the “Delete” button, and it works not on selected rows but on the window's entire scope.

    What writes pairs to the directory

    Records in the directory are created by four different parts of Elbuz.

    What writes to the directory
    Record sourceWhat the “Match type” column shows“Matched via AI” checkbox
    An operator confirmed the pair in the product matching window“Search for manually matched products”no
    Model matching (AI matching)“Search for manually matched products”yes
    The “Save automatically found products as manually matched” buttonthe name of the rule that found the pairno
    Importing pairs from a CSV file“Search for manually matched products”no

    A pair picked by the model is labeled as manual. Elbuz gives it the same match type as an operator's pair — “Search for manually matched products”. So this column can't tell an AI link from a manual one at all; the only sign is the separate “Matched via AI” checkbox.

    After that, the record lives on its own. On every price list import, the first matching rule is usually “Search for manually matched products” — it goes through the directory and puts the pairs back. That's why you never have to confirm the same item again, however many times the price list is reloaded.

    The life of one pair: at any moment it's in exactly one state, and each arrow names the event that moves it to the next
    pair confirmed by anoperator, picked by themodel,saved by the button orimported from CSVthe price list has no rowwith the same product hashthe row is back in the pricelist“Delete” — for all price listsat once,the catalog price stays old“Delete”Link appliedprice and stock go to the cardLost linkrecord alive, price and stock empty
    • Start
    • State
    • End

    The directory is shared, not per price list. Pairs are restored by product hash, not by price list number. So a pair confirmed for one supplier also picks up the same product in another supplier's price list — if the name and SKU match. The reverse holds too: breaking one row deletes the record for all price lists at once.

    If the matching rule has the flags for matching by manufacturer SKU, supplier SKU or name turned on, pairs are restored by those as well.

    The “AI matching is running for this price list” banner

    During background model matching of the price list, a blue banner appears at the top: the heading “AI matching is running for this price list”, a “Processed” line with processed and total items in percent, the warning “Manual changes may conflict with AI results right now.” and an “Open AI window” link. It rechecks the status every ten seconds.

    When the job is first detected, the message “Heads up: AI matching is running for this price list” pops up, and when it ends — “AI matching completed”. While the banner is shown, don't break links in this price list: the model is writing to the same directory right now.

    What a lost link means

    A pair stays in the directory even when its price list row is gone: the item disappeared from the file, its name changed, or the price list was reloaded. Such a record is called a lost link, and the list marks it with the “Link to price list lost” checkbox.

    The “Show lost links” toggle on the toolbar reloads the list with a different view. Off — only pairs with a live price list row. On — all lost ones as well.

    Grid with Show lost links switched on and lost rows listed
    1Show lost links switched on
    2Lost-link checkmark
    3Empty price
    1. With the toggle on, rows with lost links appear in the grid.
    2. The checkmark means the price list row no longer exists.
    3. Price stays empty because only the current price list provides it.

    A lost link has three empty columns — “Price”, “Stock status (from price list)” and “Category”: only a current price list row provides them. Names, SKUs, manufacturer and supplier in these rows come from the directory archive, and the match type comes from the record itself, the way the match was originally saved.

    The pair itself still works. As soon as the product shows up in the price list again with the same name and SKU, the link applies by itself and the card gets a price again. A lost link isn't a fault to fix — it's the normal state for a seasonal product or an item the supplier temporarily doesn't carry.

    Do you need to clean up the directory?

    The directory almost only grows. Only one thing is removed automatically — records for products no longer in the catalog: the daily cleanup deletes them. Nothing is cleaned by record age or when a price list or supplier is deleted. The other ways to shrink the table are yours: the trash icon in a row and the “Delete” button.

    This does less harm than it seems, for the same reason — the hash: a record from a dead price list will work if the same product arrives in another price list with the same name and SKU. So there's no need to clean the directory “just to tidy up” — you'll remove working pairs. What's worth knowing about leftovers:

    • Records of deleted price lists and suppliers remain. You see them only in all price lists mode, with the name from the directory archive in the “Supplier” column.
    • Records for deleted catalog products are removed by the daily cleanup. They appear when a catalog update removes a product by relevance days: it cleans the catalog but leaves the directory alone until the next cleanup. The window never shows them, so there's no need to hunt for them.
    • Directory size doesn't slow anything down. Hundreds of thousands of pairs is a normal working state.

    How to break one product's link to a price list

    This is the main reason people come here: a product card got someone else's price and stock, and the pair has to go.

    Use the trash icon in the first column of the row. Elbuz asks “Delete entry?”, after you confirm shows “Entry deleted” and removes the row from the list. There's no second question and no Undo button.

    Confirmation dialog for deleting a match record
    1Delete entry? question
    2Delete confirmation button
    1. The dialog asks whether to delete the entry.
    2. Delete confirms and removes the match immediately; Cancel closes the dialog.

    What actually happens:

    • the record is deleted from the directory — for all price lists at once, the price list number plays no part in this operation;
    • the price list row loses its “product found” flag;
    • the catalog product's link fields are cleared: supplier, price list row, price list number, sheet.

    Breaking a link doesn't bring back the old price. Only the link fields are cleared. Price, stock and quantity in the card stay as they last came from the price list — the very values you're breaking the pair over. You'll have to fix them by hand or link the product to the right price list row.

    The second, less obvious effect: the product no longer has a link to a price list row. If “relevance days” are set in the base catalog update settings, the product will eventually be disabled or deleted — depending on which of the two fields is filled in. Both are off by default, but check before, not after.

    Whether the break is permanent depends on where the pair came from.

    What happens on the next price list import
    What you brokeResult
    A manual link (or a link created by AI)Won't come back: there's no record in the directory anymore, nothing to restore
    An automatic linkWill come back if the rule that found it is still on and the data hasn't changed. To keep it from coming back, change the matching rules or the ignore lists

    Where else a link gets broken

    Three other places in Elbuz break links the same way, and they're worth knowing about because the result is the same:

    • the price list row details panel — in the price list product list;
    • the catalog product card, list of links to price lists — deleting a grid row, message “Product deleted”;
    • the “Price list products” grid — clearing the “Found” checkbox in a row.

    The third way is the bluntest: clearing the “Found” checkbox in the price list products grid deletes directory records by product hash and also by matching name, regardless of price list. One click can wipe out other suppliers' confirmed pairs for a product with the same name. To break a single pair, use the trash icon in the directory.

    How to delete all links of a supplier or the whole directory

    The first toolbar button is the only one with a label. It looks like a regular “Delete”, and its tooltip spells it out: “Delete all product matches”. The question before it runs is “Delete all product match records?”, and when it's done — “Entries deleted”.

    A single click makes Elbuz do six things:

    1. clears the “product found” flag in every price list row that has a pair in the directory;
    2. clears the link fields of catalog products mentioned in the directory;
    3. also clears the same fields in every catalog product that has no row in any price list at all — this step always runs, whatever scope you chose;
    4. deletes directory records within the scope: a price list, a supplier, or the whole shop;
    5. writes a line to the work log — “Matching directory cleared:” and the scope name, without a list or count of removed pairs;
    6. recalculates the price list counters.

    Step three is why “Delete” is more dangerous than it looks, even in single price list scope: products that never came in any price list also lose their link to a supplier.

    The scope is set by how the window was opened, not by a button. Step three comes before the fork: it runs outside the scope and is the same in all three cases
    cancelled“Delete”from the source headerfrom the supplier tabfrom the “All sources” node“Delete” button“Delete all product match records?”Nothing changesPrice list rows within the scope losethe “product found” flagLink fields cleared for catalogproducts from the directory withinthe scopeOutside the scope — the samefields cleared for all products with norow in any price listHow the window was openedRecords of this price list deletedSupplier's records deleted across allits price listsThe shop's whole directory deleted“Matching directory cleared:” line inthe work log, counters recalculated
    • Action or result
    • Decision: the path depends on the answer

    Before you click “Delete”

    What the wand and floppy disk buttons do

    Toolbar buttons two and three are bare icons, a wand and a floppy disk, named only by tooltips. Both ask first — read the question to the end: when the window wasn't opened from a source header, it starts with “The window is open without a selected price list” and warns that the operation will run for all price lists.

    “Start product matching”

    The wand removes all links in the window's scope and rebuilds them with the enabled rules; manual pairs return only if the price list has the “Search for manually matched products” rule on. The directory itself doesn't change. Restarting without losses and stopping a run across all price lists — How to restart price list matching without losing manual pairs.

    What the wand does after you confirm: the upper fork is how many price lists the run covers, the lower one is whether manual pairs come back. The directory doesn't change on any branch
    noyesfrom the source headerwithout a selected price listyesnoWand “Start product matching”Question confirmedNothing changesHow the window was openedThis one price listEach active price list of activesuppliers in turnAll links of the price list to catalogproducts removedEnabled matching rules run in theirorder“Search for manually matchedproducts” is among the enabledonesManual pairs restored by hash inrows not taken by earlier rulesManual pairs not restored, directoryrecords intactPrice list counters recalculated
    • Action or result
    • Decision: the path depends on the answer

    “Save automatically found products as manually matched”

    The floppy disk writes pairs Elbuz found by itself with matching rules to the directory as confirmed; pairs already there aren't added again. After that, a pair survives price list reloads and rule changes, but a wrong pair becomes permanent too — only the trash icon removes it. More: how to lock in good automatic pairs.

    In the “Match type” column, these records show the name of the rule that found them, not “Search for manually matched products”.

    Scenarios

    A product card has someone else's price — remove one wrong pair

    What to do. Open the directory from the header of the source the price came from: arrow next to the “Product Matching” button → “Matched products”. Find the row by filtering the “Product name from base catalog” column, click the trash icon in the first column and confirm “Delete entry?”.

    What you get. The record disappears from the directory for all price lists, the price list row loses its “product found” flag, and the catalog product's link fields are cleared. The work log gets a “Product link removed” line with the product .

    What won't happen. Price and stock in the card won't go back to the previous values — you fix them by hand. If the row's “Match type” column showed a rule name rather than “Search for manually matched products”, the pair was automatic: on the next price list import the rule will find it again.

    One supplier's pairs were bulk-matched wrong — start over

    What to do. “Directories” → “Processing price lists” → “List of price lists” → supplier → “All products” → “Additional Operations” → “Matched products”. First “Export matches to XLSX file” (down arrow), then “Delete”. The dialog must show only the red line “All links of the selected supplier will be deleted — across all of its price lists. The operation cannot be undone.”

    What you get. The file holds this supplier's directory records — exactly the ones about to be deleted. After you confirm, the supplier's records disappear from the directory across all its price lists, and the work log gets “Matching directory cleared: supplier #” with its number. Then you create the pairs again in the product matching window.

    What won't happen. No rollback — the file is your only copy. Records without a supplier, e.g. imported from a file without the “Supplier (name/address)” column, show in this list, but the cleanup skips them. Catalog prices don't change. And remember step three of the cleanup: catalog products without a row in any price list lose their link fields too.

    Troubleshooting

    Troubleshooting
    What you seeWhyWhat to do
    An “Access denied” image instead of the window, with no explanation No access to the “Matched products directory” window Ask an administrator to grant the right to your group (who has access)
    The list is empty, though pairs definitely exist Either this price list has no live rows with pairs, or the pairs were imported from a CSV file and have no price list number Turn on the “Show lost links” toggle; for pairs imported from a file, open the directory from the supplier's “All products” tab: it shows both its records and records without a supplier (import from CSV)
    The list shows products from a supplier other than the one you picked The directory was opened from the top node “All sources”: in this scope it shows records of all suppliers Close the window and open the directory from the right supplier's heading or the source header (three scopes)
    The directory has the row, but no price or stock in it This is a : there's no live price list row, so nothing provides a price Check whether the item still comes in the price list. When the product returns, the link applies by itself (lost links)
    You broke the link, but the card still has the old price Breaking a link removes the link but doesn't roll back the price, stock and quantity already copied to the catalog Fix the card by hand or link the product to the right price list row (breaking a link)
    The catalog product became disabled or disappeared after the break The product has no link to a price list row left, and relevance days are set in the catalog update settings Check both relevance days fields in the catalog update settings (breaking a link)
    You broke the link, but it came back after the price list import The link was automatic: the matching rule is still on and found the same pair again. The of such a row names the rule Change the matching rules or the ignore lists (breaking a link)
    “Processing data, please wait” and the message doesn't go away A matching run is in progress; in all price lists mode it goes through every active price list Wait; a run across all price lists is stopped with “Stop import” (restart across all price lists)
    “Operation failed!” with an error text The request to the server failed: connection dropped or timed out Try again. If it keeps happening, check the work log
    A browser system dialog “Status: … Error: …” The request failed while deleting a row with the trash icon Refresh the window and check whether the pair is gone: it may not have been deleted
    “Heads up: AI matching is running for this price list” Model matching is running in the background for the price list, and it writes to the same directory Wait for the “AI matching completed” message and reload the list (AI banner)

    FAQ

    Why does the “Match type” column say “Search for manually matched products” almost everywhere?

    That's the label for all confirmed pairs — by an operator, by the model, and imported from a file. Only pairs locked in with the floppy disk show a rule name. An AI pair is told apart only by the “Matched via AI” checkbox (what writes pairs).

    Can I restore a deleted pair?

    No: neither the trash icon nor “Delete” has a rollback, and the work log keeps only a record of the event. An automatic pair comes back on the next import if the rule is still on; a manual one comes back only from a backup file.

    I accidentally deleted all of a supplier's links. What should I do?

    If you have an XLSX file exported before the cleanup, import it back (scenario). If not, automatic pairs are rebuilt on price list import or on a matching restart, and manual ones have to be made again.

    What happens to the directory if I delete a price list or a supplier?

    The records stay, and the pair still works: if the same product arrives in another price list with the same name and SKU, the link applies. These records are visible only in all price lists mode.

    How do I delete one supplier's links without touching the others?

    Open the directory from this supplier's “All products” tab and click “Delete”. The dialog's red line must name this supplier; if it mentions “ALL links in the directory”, you opened it from “All sources” — don't confirm (operation scopes).

    Where to find it and who has access

    From the source header
    “Directories” → “Processing price lists” → “List of price lists” → pick a source → arrow next to the “Product Matching” button → “Matched products”
    From the supplier tab
    the same list → click the supplier heading → “All products” tab → “Additional Operations” menu → “Matched products”
    From the “All sources” tab
    the same list → top node “All sources” → “All products” tab → “Additional Operations” menu → “Matched products”
    Access right
    the “Matched products directory” window; the trash icon and “Delete” can't be restricted separately
    Import from CSV
    also requires access to the “Product matching” window
    Prepare
    an imported price list and at least one confirmed pair

    The directory has no menu item and no desktop shortcut: it opens only from the price list module. Without access to the window, you see an “Access denied” image instead of its content.

    Price list card header with the Product Matching button and its opened menu
    1Product Matching button
    2Arrow that opens the menu
    3Matched products item
    1. Product Matching — the main button of the card header.
    2. The small arrow to the right of it opens a menu of five items.
    3. Matched products — opens the directory window.

    In all three cases the item has the same name and opens a window that looks the same. But which one you used decides which records the buttons inside will work on.

    Related topics

    Was this article helpful?

    Missing a feature? Tell us.

    We will add it for free if other users need it too.

    If only you need it, we will agree on it separately.

    Article contents

    Suggest an improvement

    Try Elbuz free