Records

Records are the rows you enter into your tables. They are the data itself — the thing everything else in Grunddata exists to protect.

The shape of a record is decided by the table’s attributes.

What a record consists of

Part	Description
source_id	The permanent identity of the row. See Identifiers.
version	Which revision this is. Increments on every change.
Code	The built-in business key. Required on MANUAL tables.
Name	The built-in readable label. Required on MANUAL tables.
Attribute values	One value per attribute you defined.
is_deleted	Whether this version represents a deletion.
Created by / created at	Who wrote this version, and when.

Ways to maintain records

Excel plugin

The normal way to work, and what most business users will use. Download a table, edit it like a spreadsheet, upload with a commit message. See Excel plugin.

Google Sheets

The same workflow for Google Workspace teams. See Google Sheets add-on.

Web app

Best for single rows — a quick correction, or adding one new lookup value. Not suited to bulk work.

API

For IT and ETL. The same API the web app and the plugins use, so the same validation applies.

Pick the right tool for the size of the job

One row: the web app. Twenty rows: the web app is still fine. Two hundred rows: use the Excel plugin or Google Sheets. Two hundred thousand rows on a schedule: use the API.

The records grid

Open a table and scroll to Records. The grid shows:

Column	Shown when
Source ID	Always
Code, Name	Only on MANUAL tables
One column per attribute	Always, in your saved attribute order
Version	Always
Deleted	Always

Paging is server-side, so the grid stays fast on large tables. Choose 5, 10, 20 or 50 rows per page.

Domain columns show the readable label

For a domain attribute the grid shows the pretty value — the referenced record’s Code - Name — not the stored source_id. The source_id is what is actually stored.

Add a record

1. Open the table and click Add Record.

2. On a MANUAL table, fill in Code and Name. Both are mandatory. On an AUTO table these fields are not shown.

3. Fill in the attributes. The editor adapts to the datatype:

   • date and datetime get a date picker.
   • boolean gets a checkbox.
   • Domain attributes get a search picker — click the magnifying glass, search by code or name, and select.

4. Save.

The domain picker searches Code and Name

Type part of either. Searching nord will find NORDEASE - Nordea Bank Abp. The list is paged, so a large lookup stays responsive.

If the save is rejected

The dialog stays open and shows the validation message. Nothing is written. Correct the value and save again — see Validations and Record audit.

Edit a record

Click a row in the grid, or the pencil icon. Change what you need and save.

Editing does not overwrite the existing row. It appends a new version with version + 1. The previous version stays in history. See Commits and versions.

See a record’s history

Open a record and select History to see every saved version of that row. Each entry shows the value at that point, the version number, who made the change and the commit that grouped it with related work.

Use history to answer a specific question

The table overview’s Recent commits panel is best for “what changed in this table?”. Record history is best for “when did this particular value change, and who changed it?”.

Delete a record

Click the trash icon on the row and confirm.

Deletes are soft

Deleting appends a new version flagged as deleted. The record stops appearing in the grid, in downloads and in OData, but the history remains intact and attributable.

Do not delete records that are referenced

If the record is a value in a domain table, other records may point at it. Check before deleting, or you will leave references to an inactive value.

How many records can I store?

There is no limit imposed by Grunddata. Listings and the OData feed are paged, and filtering and ordering are pushed down to the database.

Excel has its own limits

Grunddata will happily hold millions of rows, but Excel will not enjoy downloading them all at once. For very large tables, work in slices or use the OData feed with a $filter instead of a full download.

Example

A Reports table:

Report name	Owner	Recipient	Frequency
Counterparty limit follow up	Credit Risk	Client executives	Daily
Value at risk report	Market risk	CFO, CRO	Weekly
Internal capital requirements	Risk	CEO, CFO, CRO	Monthly

Tip

Owner, Recipient and Frequency above are all good candidates for domain tables. Free text here means you will eventually have Market risk, Market Risk and market-risk all in the same column.