# Computed Fields FAQ

Source: https://userguide.qualytics.io/fields/computed-fields/faq/

## About Computed Fields

#### What is a Computed Field?

A Computed Field is a derived field attached to a container. The definition (transformation and source fields it reads) lives with the container's metadata; the value itself is produced when Qualytics profiles or scans the container. The Computed Field appears alongside the container's native fields and can be profiled and monitored the same way.

#### How is a Computed Field different from a regular field?

Regular fields come from the container's schema: they reflect what the source system actually stores. Computed Fields are declared by you and evaluated by Qualytics on top of that schema. They do not exist in the source system, and deleting a Computed Field never touches the source data.

#### When should I use a Computed Field instead of a quality check filter?

Use a Computed Field when the derived value needs to be visible on the container, referenced by multiple quality checks, or reported on. Use a check filter when the logic is one-off and only relevant to a single check.

## Transformation Choice

#### Which transformation type should I pick?

Reach for the specialized transformations first: **Cast** for data-type conversion, **Cleaned Entity Name** for stripping business signifiers, and **Convert Formatted Numeric** for formatted-string amounts. Reserve **Custom Expression** for logic that spans multiple fields, uses conditional branches, or requires window functions.

#### Can I chain Computed Fields?

Yes. A Computed Field can use another Computed Field as its input, for example a Custom Expression `lower(name_upper)` over a field `name_upper`. What happens when you delete the field it uses depends on how the dependency is declared. A Computed Field that lists it as a source field (for example, a Cast) is deleted along with it. A **Custom Expression** that references it in its expression is left in place, and the container's next profile fails until you delete or edit that Custom Expression, so delete it first.

#### Does Custom Expression support any Spark SQL function?

Almost. Custom Expression accepts any valid Spark SQL expression that produces one value per row against the container's fields, including window functions (`LEAD`, `LAG`, `ROW_NUMBER`) and conditional constructs (`CASE`). Aggregate functions (`SUM`, `AVG`, `COUNT`) work only as window functions with an `OVER` clause, for example `SUM(amount) OVER (PARTITION BY customer_id)`. A bare `SUM(amount)` fails validation.

## Behavior

#### When is a Computed Field's value computed?

At Profile and Scan. The transformation runs against the sample the platform reads during Profile, and against the rows Scan pulls to evaluate quality checks. There is no persistent copy of the derived value between runs.

#### Does saving a Computed Field trigger a profile?

Yes. Saving runs a quick profile of the container right away, which validates the definition and computes the field's first values, so they appear within seconds. The profile is listed in the container's Activity.

#### What happens when a source field is dropped from the source?

The Computed Field's next profile marks the derived value as Missing. The Computed Field definition itself is preserved, so restoring the source field brings the Computed Field back on the next profile.

#### What happens when a source field is excluded via Qualytics?

For a Computed Field that declares the field as a source field, the output is also excluded, but the definition is preserved. A **Custom Expression** computed field declares no source fields, so the cascade does not reach it: it stays Active when a field its expression uses is excluded, and keeps computing from that field's values. For a Computed Field that was excluded, restoring the source field does not restore it automatically: restore the source field first, then restore the Computed Field itself. This is different from deleting a Computed Field, which permanently removes both the definition and the output.

## Permissions

#### Who can create a Computed Field?

Any user with the **Editor** team permission on the parent datastore. An **Author** on the datastore can also create a Computed Field when `owner_id` is omitted or set to their own user id (the Hybrid Gate). Reassigning ownership after creation always requires the **Editor** team permission.

#### Who can view a Computed Field?

Any user with the **Reporter** team permission or above on the parent datastore.

## Troubleshooting

#### The Custom Expression fails validation, what should I check?

Confirm every source field referenced in the expression exists on the container by its exact name. Check for typos in function names. Malformed CASE, missing END, unbalanced parentheses, or an aggregate function used without an `OVER` clause are also common causes.

#### My Computed Field shows Missing after a source update, what should I do?

Open the field's profile page and confirm which source field is missing. If the source field was renamed at the source, edit the Computed Field to reference the new name. If the source field was dropped and needs to come back, restore it at the source and wait for the next Sync and Profile.

#### Can I recover a deleted Computed Field?

No. Deletion is irreversible. The definition, the derived output, and the quality checks that targeted the field are removed together. Expected Schema checks are the exception: they stay and drop the field from their list. Copy the transformation before you delete so you can recreate it later.

**Still have questions?**
Contact the Qualytics support team at
[support@qualytics.ai](mailto:support@qualytics.ai).

*Last updated: October 20, 2026.*