Ensuring Data Validity in PI Vision: A Guide for Manual Data Entry

Manual data remains necessary in many industrial workflows. Examples include laboratory results, field readings, inspection values, production adjustments, and approved corrections.

Manual data can be reliable when the entry process uses clear validation and traceability controls.

Define the authoritative entry process

Users should enter governed data through an approved tool or workflow.

Avoid uncontrolled spreadsheets, direct database changes, and informal copy-and-paste processes that bypass validation.

The approved process should define where the value is written and who can write it.

Validate the value before write

Use checks that match the measurement.

Common checks include:

• Data type

• Minimum and maximum range

• Required precision

• Engineering units

• Required fields

• Allowed codes or states

A generic range is not sufficient for every asset. Use process-specific limits when the risk justifies them.

Validate the timestamp

Timestamp errors can be difficult to detect after the value is stored.

Confirm:

• Date and time

• Time zone when relevant

• Whether the entry represents sample time or entry time

• Whether a value already exists at that timestamp

Do not silently overwrite an existing historical value.

Capture context for important entries

For corrections and regulated workflows, record enough context to explain the change later.

This can include:

• User

• Entry time

• Effective timestamp

• Reason code

• Comment

• Work order or laboratory reference

The required metadata depends on the business process and compliance requirements.

Prevent unit ambiguity

A value without a clear unit can be dangerous.

Display the expected engineering unit in the entry form. Convert units through a controlled rule if conversion is allowed. Do not rely on the user to remember the expected unit.

Handle corrections as controlled changes

When a historical value must be corrected, preserve the required audit evidence.

The user should know whether the process replaces the value, adds a corrected event, or keeps an external audit record. The correct method depends on the application and regulatory context.

Monitor manual data after entry

Manual values should not bypass normal data-quality checks.

Review stale behavior, unexpected values, duplicates, and downstream use in the same way that you review automated data.

The objective

Manual entry is not inherently lower quality than automated collection. The risk comes from uncontrolled entry.

A reliable process validates the value and timestamp, records required context, controls corrections, and makes the result traceable to its source.