Attributes

Attributes are the columns of a table. They define what kind of data each field holds and what counts as a valid value. You can create any number of them.

Every table has built-in identifier and display name attributes. Standard (AUTO) tables receive them as Code and Name; reference tables can choose their names when created. You do not create or reassign these roles from the attribute editor. See Identifiers.

Structure

Field	Description	Required
Name	The attribute name. This is the column header users see in Excel and Google Sheets.	Yes
Datatype	One of the supported datatypes below	Yes (unless a domain table is chosen)
Required	Makes the field mandatory on new records	No
Unique	No two records may share the same value	No
Max length	Maximum characters, string only	No
Precision / Scale	Total digits and decimal digits, decimal only	No
Domain table	Reference records in another table	No
Cascade parent	Narrow the domain values based on another attribute	No
Validation rules	JSON Schema rules for the attribute	No
Excel formula	A spreadsheet expression calculated from other attributes	No

Datatypes

Datatype	Stores	Extra settings
string	Text	Max length
integer	Whole numbers	—
decimal	Fixed-point numbers	Precision (total digits), Scale (digits after the point)
boolean	True / false	—
date	A calendar date (yyyy-MM-dd)	—
datetime	A date and time (yyyy-MM-dd HH:mm)	—

Pick decimal, not string, for numbers

It is tempting to store an amount as text because “it came from a spreadsheet that way”. Don’t. decimal with an explicit precision and scale gives you numeric validation, comparison rules and correct sorting in Power BI. A 2 in a text column sorts after 10.

Domain table attributes

An attribute can point at another table instead of holding a raw value. The referenced table becomes a controlled list of allowed values.

1. In the attribute editor, choose the domain table model.
2. Choose the domain table. Only MANUAL tables appear in the list.
3. The datatype becomes read-only — it is determined by the referenced table.

When a user fills in the field they get a searchable picker showing Code - Name from the target table. Grunddata stores the target record’s source_id, and resolves the readable label on the fly.

Reference instead of validating

If you find yourself writing a One of (list) validation with more than a handful of values, you probably want a domain table instead. A domain table can be maintained by the business without a structure change, and it shows up in the picker with readable labels.

Full details in Domain tables.

Cascading domain attributes

A domain attribute can be made to depend on another domain attribute on the same table. Tick Cascade domain values from another domain attribute and choose the cascade parent attribute.

The effect: when a user picks a value in the child attribute, the candidate list is filtered by whatever they selected in the parent.

Typical use

country → region → branch. Selecting Sweden in country means the region picker only offers Swedish regions. This turns a 2,000-row lookup into a usable one.

Attribute order

The order of attributes controls the column order everywhere: the records grid, the Excel download and the Google Sheets download.

Use the up and down arrows on the attributes page to rearrange, then save the order.

Caution

If you navigate away with unsaved order changes, the app will warn you. The new order is not applied until you save.

Create an attribute

1. Open the table and click Attributes.
2. Click Add attribute.
3. Fill in the name, datatype and constraints — or select a domain table.
4. Optionally add validation rules.
5. Save.

Update an attribute

On the attributes page, click the pencil icon next to the attribute.

Changes that are not retroactive

Grunddata deliberately does not rewrite historical data when you tighten a rule:

• Required applies only to records created or updated after the change. Existing rows keep their empty values.
• Max length applies only to new values. Existing longer values are left alone.

If you need existing data to conform, fix the data — for example by downloading the table in Excel, correcting it, and uploading with a commit message explaining the clean-up.

Changes that can be blocked

• Unique cannot be enabled if the table already contains duplicate values for that attribute. Clean the duplicates first.
• Unique cannot be disabled on the identifier attribute of a reference table — the identifier role requires it. The checkbox is disabled for that attribute.
• A domain table cannot be set on an attribute in a table that already holds records. Existing values could not be resolved to target records, so the change is refused.

Renaming an attribute

Renaming is safe internally — records and views reference the attribute by source_id. But the name is the column header in Excel and Google Sheets, so tell your users before you change it, and check any view field names you may also want to keep aligned.

Excel formulas

An attribute can hold an Excel formula as well as the calculated values that are saved in its records. This is useful when a worksheet calculates a value, for example:

={{Balances.Closing}}-{{Balances.Opening}}

Grunddata stores stable references to the source table and attributes behind that readable expression. Renaming Closing later does not break the relationship: the stored identifiers stay the same, and the displayed formula follows the new names.

Formulas come from Excel

For now, formulas are created by the Excel plugin when it creates a new table. The web app shows the expression in the attribute editor as read-only, so you can see how a calculated column is defined without accidentally changing it.

Calculated value and formula are both useful

The formula explains how Excel arrived at a number. The calculated value is what Grunddata saves in the record, so the data remains useful to the web app, API and OData consumers that do not run Excel calculations.

Delete an attribute

On the attributes page, click the trash icon next to the attribute.

Danger

Deleting an attribute removes that column from the table’s current structure. Do it deliberately, and check first whether any view exposes it.

A formula dependency cannot be deleted

Grunddata blocks deletion of an attribute or table that is used by an attribute formula. Remove or change the formula first. This prevents a calculated column from being left with a broken reference.

Attribute names are unique within a table

You cannot create or rename two current attributes to the same name on one table. It keeps Excel and Google Sheets headers unambiguous.

Examples

An accounts table:

account_number	account_name	account_type
1200	Nominal amount	NOM
1300	Market value adjustment	MTM
1400	Interest rate	INT

An instrument hierarchy:

Instrument type level 1	Instrument type level 2	Instrument type level 3
Derivative	Interest rate swap	Interest rate swap
Derivative	Options	Equity option
Derivative	Foreign exchange swap	Foreign exchange swap
Security	Bond	Covered bond
Security	Bond	Zero coupon
Loan	Open ended	Open ended
Deposit	Deposit	Deposit

Tip

The hierarchy above is a good candidate for cascading domain attributes: make each level a domain attribute whose parent is the level above it.