Organize information

Audit Excel headers, field definitions, and data types

Compare a workbook with its intended schema before import, analysis, or automation

7 min setupTested with:ChatGPTReviewed: 2026-08-28

Add your context

Your text stays in this browser. AILesson Prompts does not send it to a model or server.

Workbook purpose and consumerExpected schemaObserved headers and sample values

Your prompt

Unfilled fields remain visible as placeholders, so you can still copy and edit the prompt

Audit an Excel workbook against its intended schema.

Purpose and downstream consumer:
[purpose]

Expected schema:
[expected]

Observed headers and representative values:
[observed]

Check header names, order only when order matters, missing and unexpected fields, data types, date and number storage, requiredness, uniqueness, allowed values, and cross-field rules. Distinguish cell formatting from underlying value type. Return a data dictionary, schema-difference table with severity, row-level validation formulas, blocking issues, and an acceptance checklist. Do not declare conformance from display appearance alone.

Private by defaultPrompt assembly happens locally in your browser. Avoid placing confidential information into any AI service unless your organization allows it.

From input to outcome

A worked example

See how concrete context turns this recipe into a usable result

Actual input

Workbook purpose and consumer: Upload weekly orders to an operations system. Column order is irrelevant, but exact header names and field types are required.
Expected schema: OrderID required unique text; OrderDate required Excel date; CustomerID required text; Amount required number >=0; Currency required one of USD, EUR, GBP; Cancelled optional TRUE/FALSE. No extra fields containing notes may be uploaded.
Observed headers and sample values: Headers: Order ID, OrderDate, CustomerID, Amount, Currency, Cancelled, InternalNotes. Samples include Order ID 00125; OrderDate displayed 08/27/2026 but imported as text in some rows; Amount '$1,240.00' stored as text; Currency usd; Cancelled Yes.

Example output

Schema differences

Issue	Severity	Evidence	Required action
Order ID header differs from OrderID	Blocking	observed space	rename exactly
InternalNotes is unexpected	Blocking/privacy	extra sensitive-capable field	exclude from upload
OrderDate mixed type	Blocking	some values are text	convert and validate, do not only reformat
Amount stored as currency text	Blocking	$ and comma in text	parse under confirmed locale, retain raw
Currency uses lowercase	Fixable	usd	map approved case to USD
Cancelled uses Yes	Blocking until mapped	contract requires Boolean	approve Yes/No mapping or reject

Create a data dictionary with field, definition, stored type, required, allowed values, example, and validation. Row checks can include =AND(ISTEXT([@OrderID]),[@OrderID]<>""), =ISNUMBER([@OrderDate]), =AND(ISNUMBER([@Amount]),[@Amount]>=0), =ISNUMBER(XMATCH([@Currency],{"USD","EUR","GBP"})), and =OR([@Cancelled]=TRUE,[@Cancelled]=FALSE,[@Cancelled]=""). Check uniqueness with COUNTIF. Accept only when exact required headers exist, InternalNotes is excluded, every row passes type and domain checks, OrderID count equals unique OrderID count, and rejected rows are listed separately. Preserve 00125 as text.

Why this works

1
A schema contract catches structural failures before they become misleading analysis results.
2
Testing underlying value types reveals dates and numbers that only look correctly formatted.

Check the result

Are required fields, keys, and allowed values explicitly documented?
Do validation checks inspect stored types rather than display formats?
Are blocking import failures separated from nonblocking presentation differences?

More ways to explore