Catalog Admin
Setup Guide — Curating the Vendor Metadata Catalog Behind Discovery
Module: Discovery Module › Catalog Admin | Last updated: August 2026
Contents
- Overview
- Finding Your Way Around
- The Overview Tab
- Adding a Vendor Dictionary
- Working With a Vendor's Files
- AI Descriptions
- Deriving Relationships (Foreign Keys)
- Industry Packs
- Live Harvest
- Business Object Names
- Load History
- Quick Reference
1. Overview
Catalog Admin is where you manage the vendor metadata catalog — the library of vendor "dictionaries" that describe the tables and columns of source systems like SAP, Oracle EBS or Infor LN. Each dictionary lists a vendor's business objects, their tables, the columns in those tables, and how those tables relate to one another.
This catalog is what the Discovery Workbench reads when you profile a source in Catalog mode: instead of looking at raw, cryptic table and column names, Discovery can show a business object's friendly name, a column's description, and its declared keys and relationships. A richer catalog means a clearer, faster Discovery experience for everyone who uses it.
What you can do here
- Add a vendor dictionary by uploading its source file(s).
- Manage a vendor's files — add more, remove them, or export them.
- Fill in missing descriptions using AI.
- Work out relationships (foreign keys) between tables automatically.
- Pull metadata straight from a live system with Live Harvest.
- Give business objects friendly names that show up across OnCoor.
WHERE TO FIND IT — Catalog Admin opens from its own left-nav entry. The page is organised into three tabs: Overview, Live Harvest, and Business Object Names.
YOU NEED THE "MANAGE CATALOG" PERMISSION — Anyone can open Catalog Admin to look, but adding vendors, enriching descriptions, deriving relationships, and adding or removing files all require the Manage Catalog permission. Without it, those buttons are visible but greyed out, and a banner reminds you that you have read-only access.
2. Finding Your Way Around
The page has three tabs across the top:
| Tab | What it's for |
|---|---|
| Overview | The main workspace — see what's in the catalog, add vendor dictionaries, manage their files, run AI descriptions, and derive relationships. |
| Live Harvest | Connect to a live source system and read its metadata directly into the catalog, rather than uploading a file. |
| Business Object Names | Give the raw grouping keys friendly display names that appear throughout OnCoor. |
Most day-to-day work happens on the Overview tab, so that's where the rest of this guide starts.
3. The Overview Tab
At the top of the Overview tab you'll see four summary cards that give you the state of the whole catalog at a glance:
| Card | What it shows |
|---|---|
| Vendors | How many distinct vendor systems are in the catalog. |
| Versions | How many vendor/version combinations exist (a vendor can have several versions). |
| Objects | The total number of business objects across the whole catalog. |
| Enriched | The percentage of columns that have an AI-written description. |
Below the cards are the main working areas: the Add a vendor dictionary form, the Catalog inventory table, and the Load history log. Each is covered in its own section below.
4. Adding a Vendor Dictionary
A vendor dictionary describes one version of one vendor system. You create one by uploading its source file(s) in the Add a vendor dictionary panel.
Steps
- Enter a Vendor system code — for example,
ORACLE_EBS. - Enter a Version — for example,
R12_2. - Optionally add a Description to explain what this dictionary is.
- Click Choose files and pick one or more Excel files.
- Click Add vendor.
When it finishes, a message confirms how many objects, attributes and relationships were loaded.
CODE RULES — Vendor system and version codes may use letters, digits,
_and-, and must start with a letter or digit. The vendor/version pair is how the rest of OnCoor identifies this dictionary, so pick something stable and meaningful.
FILE FORMAT — Source files must be Excel (
.xlsxor.xls) in OnCoor's standard dictionary layout. Every row needs at least a Business Object Name and an Object Type, plus the table and column details (Table Name, Column Name, Attribute Name, Attribute Description, Col Data Type and Col Scale). Decimal columns also need Col Precision and Col Scale. If you're not sure of the layout, export an existing vendor (see Working With a Vendor's Files) and use it as a template.
Once a dictionary exists, you don't re-upload the whole thing to change it — instead you expand its row in the inventory and add or remove individual files.
5. Working With a Vendor's Files
The Catalog inventory table lists every vendor/version in the catalog, one row each, showing its object count and AI-description coverage. Each row is tagged so you know where it came from:
| Tag | Meaning |
|---|---|
| Bundled | Shipped with the OnCoor release. These are managed in source control and can't be changed here. |
| User-added | Uploaded at runtime through Catalog Admin. These you can fully manage. |
Click the arrow at the start of a row to expand it and reveal that vendor's source files. From here you can work with the whole vendor or a single file:
| Action | What it does |
|---|---|
| Add file(s) | Upload more source files into a user-added vendor. The catalog rebuilds to include them. |
| Export Excel | Download the vendor's dictionary as Excel — a zip with one workbook per source file (or a single workbook for one file). The download is in the standard format, so it can be re-imported. |
| Export (per file) | Download just one source file as a re-importable Excel workbook. |
| Remove (per file) | Delete a file from a user-added vendor. The catalog is rebuilt from the files that remain, so objects only that file contributed disappear. Removing the last file removes the whole vendor. |
| Promote to Catalog | For newly harvested files sitting in staging, move them into the delivered catalog directory (the permanent source of truth). |
DELIVERED FILES ARE PROTECTED — Files that ship with the release show as read-only. If you enrich or derive relationships on one, OnCoor asks you to confirm first, because you're editing source-controlled content.
REMOVING FILES REBUILDS THE CATALOG — Because the catalog mirrors the files, removing a file also removes anything only that file provided. OnCoor warns you before you remove the last file, since that deletes the entire vendor/version.
6. AI Descriptions
Source dictionaries often arrive with plenty of tables and columns but few written descriptions. Catalog Admin can fill those gaps for you: for any column that doesn't already have a description, it can generate a plain-language one using AI. Better descriptions flow straight through to Discovery, making profiling results easier to read.
The Descriptions (AI) column in the inventory shows coverage as "written / total" — for example, 1,200 / 4,000 means 1,200 columns have a description and 4,000 columns exist in total. A green tick means every eligible column is covered.
To enrich descriptions
- Click Enrich all on a vendor row to write descriptions for every column that's still missing one.
- Or expand the row and click Enrich file on a single file, to do just that file.
While it runs, a live progress bar shows how far along it is. When it finishes, a message tells you how many descriptions were written and whether any remain.
WHY ENRICH ONE FILE AT A TIME? — AI enrichment uses a daily budget. Enriching a single file lets you spend that budget deliberately, a piece at a time, rather than all at once on a large vendor.
AI ONLY FILLS BLANKS — Enrichment only writes descriptions for columns that don't already have one. Existing descriptions are never overwritten. Some file types (for example, certain raw source formats) can't be enriched in place and are marked "n/a" — re-author those as a standard Excel file if you need descriptions in them.
7. Deriving Relationships (Foreign Keys)
Relationships tell OnCoor how tables connect — which column in one table points at the key of another (a foreign key), and which columns are validated against a lookup (a check reference). Catalog Admin can work these out for you automatically by analysing the dictionary, so you don't have to map them by hand.
To derive relationships
- Click Derive FK (all) on a vendor row, or Derive FK on a single file.
- OnCoor analyses the data and opens a preview showing the foreign keys and check references it found, each with the method used and a confidence level (High or curated/Medium).
- Review the list. Every proposed relationship is selected by default; untick any you don't want, or use the filters and bulk buttons to narrow the list.
- Click Apply & reload.
Applying writes the relationships into the dictionary and reloads the catalog so Discovery immediately sees them.
The preview may also list Unresolved items — columns that look like they should link somewhere but need a person to decide, because the target was ambiguous. These are shown for your awareness and are not linked automatically.
YOUR CONFIRMED RELATIONSHIPS ARE KEPT — If a relationship was already set to a confirmed value, deriving again leaves it untouched. To deliberately replace confirmed relationships, tick Overwrite existing/confirmed relationships before applying.
8. Industry Packs
Some vendor dictionaries cover several industries at once (for example, a core set of objects plus industry-specific ones for retail, utilities, and so on). The Overview tab shows an Objects by industry strip so you can see how a dictionary breaks down.
Objects by industry — Each chip names an industry and its object count, for example Retail · 320. A plain (grey) chip is the core, cross-industry set; coloured chips are specific industries. The strip lets you see at a glance how much of a dictionary belongs to each industry.
Export from — When more than one vendor/version is loaded, the strip needs to know which one to break down. The Export from dropdown (top right of the strip) selects that source, and the industry chips update to match. This same selection is what any pack you export is authored from.
Export pack — If you have the Manage Catalog permission, each industry chip has an Export pack button. Clicking it creates a self-contained pack of just that industry's objects, taken from the vendor/version chosen in Export from. This is useful for shipping a focused, industry-specific slice of the catalog rather than the whole thing. After exporting, OnCoor tells you where the pack was written and how to install it.
Inside an expanded vendor row, a Layers strip shows the same breakdown for that one version, including any Customer delta — objects and columns added on top of the delivered baseline through Live Harvest.
THE STRIP FOLLOWS "EXPORT FROM" — The industry breakdown always reflects the vendor/version selected in Export from, not the whole catalog. Switch the dropdown to see a different source's industries and to choose which one a pack is built from.
9. Live Harvest
The Live Harvest tab reads metadata straight from a connected source system, so you can build or extend a dictionary without preparing an Excel file first. You review everything object by object, then commit only what you want into the catalog.
How it works
- Pick a vendor from the tabs (for example, an SAP system or an adapter-based source like Oracle or SQL Server).
- Choose a connection and set the Dictionary schema (see Dictionary schema). Use the test connection button to confirm the system is reachable.
- Get objects — Live Harvest lists the source's objects (for SAP, its packages) with a table count for each. You can search by name, or type a table name to find the object that contains it. Use the Object type control (see Object type) to say which kinds of tables to list.
- Load an object to see its tables in the middle panel. Expand any table to inspect its columns, primary keys, foreign keys and check references.
- Select the tables you want. Handy shortcuts select the natural boundaries for you: FK picks every table involved in a foreign-key relationship, and FK + T adds the related configuration (check) tables that live inside the same object.
- Commit to catalog to add the selected tables under the chosen business object.
Tables that are already in the catalog are locked so you don't add them twice. The right-hand panel highlights check tables the object references, flags any that still need harvesting, and lets you pull in shared configuration tables as their own objects.
Dictionary schema
This field tells Live Harvest where in the source to read the metadata from. What it means depends on the vendor:
- For an SAP system it's the Dictionary schema — where the data-dictionary tables live. The system fills in a sensible default.
- For a database source (Oracle, SQL Server) it's the DB schema (the schema or owner to read), and it's optional.
- For a file-share source it's the Directory — the subfolder of the share to read, also optional.
Object type
For SAP vendors, the Object type dropdown filters which kinds of tables are listed: Transparent tables (the default — the real data tables), Views, Structures, or All. For database and file sources this control is replaced by a Group by option instead, which decides how the source's tables (or files) are gathered into browsable objects — by DB schema, one per table, or by name prefix (and, for file shares, by directory, one per file, or by name prefix).
Commit vs Delta
Live Harvest offers two ways to finish, chosen with the Action picker:
| Action | What it does |
|---|---|
| Commit to catalog | Adds the selected tables into the target vendor/version. |
| Delta vs baseline | Compares the live system against an existing catalog version and shows only what's changed — new standard tables and columns, and custom (Z/Y) additions. Applying a delta adds those changes on top of the baseline without touching the rest. |
The delta preview labels each change clearly (for example, "New table (standard)" or "New column (custom Z/Y)"), so you can see exactly what a customer has added to a standard system before you bring it in.
DELTA IS ADD-ONLY — Applying a delta brings in new tables and columns; it doesn't remove anything from the baseline. This makes it a safe way to keep a delivered dictionary current with a customer's real system.
AVAILABLE ACTIONS DEPEND ON THE SOURCE — Not every source system supports object browsing or delta comparison. Live Harvest only shows the actions a given source can actually do.
Harvested objects
The strip along the bottom of the Live Harvest tab, Harvested objects, is your running total for the current target — it lists every business object already harvested into the selected vendor and version, with a table count for each, and a filter box to find one quickly. It updates as you commit, so you can see your work accumulate.
Two fields sit above it and set the target you're committing into:
- Vendor — the catalog identity for the source. For database and file sources you can catalog a specific system under its own name here (for example, a particular provider rather than the generic database type). For SAP the vendor is fixed.
- Version — the vendor version you're building, for example
S4_LIVE.
Once a vendor/version already contains objects, both fields lock so you can't accidentally retarget an in-progress dictionary.
Add new version — To start a fresh version of the same vendor (rather than adding to the existing one), click Add new version. This clears the version field and empties the Harvested objects strip, unlocking the field so you can type a new version name and harvest into it from a clean slate.
LOCKED FIELDS ARE A SAFEGUARD — A greyed-out Vendor or Version field means that target already has objects. That's deliberate — it stops harvested tables landing in the wrong version. Use Add new version when you genuinely want a new one.
10. Business Object Names
Source systems group their tables under keys that are often long and cryptic — for example, "Application development R/3 material master prior to 3.0". The Business Object Names tab lets you replace that with a friendly display name like "Material Master".
To add a friendly name
- Choose the vendor / version the business object belongs to.
- Pick the business object (its raw key) from the list.
- Type the Friendly name you want shown.
- Click Save.
The friendly name then appears everywhere the business object is shown — Setup, Mapping, and the Catalog Report and its downloads.
Current names
Below the form, the Current names table lists every friendly name you've already set. Each row shows the Vendor it belongs to, the Friendly name on display, and the underlying Business object key it maps to — so you can see at a glance which raw key each label stands in for. Use the Remove button on a row to delete a name; the object then falls back to showing its raw key again. If no names have been set yet, the table simply says so.
To change an existing name, add it again from the form with the new text — saving the same business object updates its name in place.
THE UNDERLYING KEY NEVER CHANGES — Friendly names are display-only. OnCoor keeps the original grouping key exactly as it is, so grouping, exports and packs are unaffected. You're only changing the label people see.
11. Load History
At the bottom of the Overview tab, the Load history table records every catalog operation — imports, file additions and removals, harvests, and so on. Each row shows when it happened, the action, the vendor/version affected, object/attribute/relationship counts, whether it succeeded or failed, and who did it. It's a useful audit trail when you want to confirm a change went through or trace who last touched a dictionary.
12. Quick Reference
A fast lookup for the most common actions.
| I want to… | Do this |
|---|---|
| See what's in the catalog | Open Catalog Admin → Overview tab |
| Add a new vendor dictionary | Overview → Add a vendor dictionary → fill in codes, choose Excel files, Add vendor |
| Add more files to a vendor | Expand its row → Add file(s) |
| Remove a file | Expand its row → Remove (removing the last file removes the vendor) |
| Download a dictionary as Excel | Expand its row → Export Excel (or Export on a single file) |
| Fill in missing column descriptions | Enrich all on the row, or Enrich file for one file |
| Work out table relationships | Derive FK → review the preview → Apply & reload |
| Read metadata from a live system | Live Harvest tab → pick connection → get objects → Commit to catalog |
| See only what a customer changed | Live Harvest → Action: Delta vs baseline |
| Give a business object a friendly name | Business Object Names tab → pick object → type a name → Save |
| Export just one industry's objects | Overview → Objects by industry → Export pack |
| Check who changed what | Overview → Load history |
Source: OnCoor Discovery Module product documentation, written for end users. For the latest screens and options, always refer to the in-app interface.