External website; opens in a new tab.

Software & Apps

What should be checked when spreadsheet records move into a database?

Reconcile meaning, relationships and permissions, not just row counts; rehearse imports and retain an independently recoverable source.

Software & Apps

Source review 21 November 2026

A spreadsheet import can finish without an error and still be wrong. Codes may lose leading zeros, dates may change interpretation, duplicate people may be merged and relationships may disappear. Before building an import, define the records and business rules with the people who use them. A database cannot decide whether two similarly named customers are the same legal organisation.

Define the mapping and exceptions

Inventory worksheets, hidden columns, formulas, attachments and downstream reports. Identify a stable record key and the meaning of each field. Decide how blank, unknown and genuinely zero values differ. Specify date formats, currency, time zones, character encoding and which values need validation. Preserve the original material in an authorised recoverable location rather than editing the only source until it matches the new schema.

Record who decides ambiguous mappings and duplicate handling. A technical script should produce a useful exception report, not silently guess. If a sheet has a formula total, decide whether to migrate its inputs, a calculated result or both with a defined purpose. Record counts alone will not reveal a calculation that changed meaning.

Rehearse without uncontrolled live data

Use synthetic or separately authorised minimised records in a test environment. Check relationships and constraints before importing the whole set. A migration needs an explicit treatment for rejected records and a way to rerun without duplication. Record source version, mapping version and completed batch/checkpoint so an interruption does not make the current state unknowable.

Agree representative reconciliation cases: number of active customers, linked orders, outstanding values, required documents and unusual records. Review examples with the business owner. A count match is useful but insufficient; two swapped amounts can leave the total unchanged. Compare meaningful invariants and a controlled sample, and retain the reason for each approved cleansing decision.

Plan the cutover and recovery

Decide when people stop editing the old system and how any final changes will be captured. Concurrent writing to two sources can create a reconciliation problem that no import script resolves by itself. Verify backup restoration, database/configuration compatibility and the rollback decision before the consequential switch. A successful backup command alone does not show the business can recover.

After import, test permissions as ordinary users, including denial across unrelated customers/projects. Check exports, search, reports and the restart/recovery path. Keep engineering verification and the customer’s acceptance separate; a developer’s “script completed” message does not establish the records are commercially correct.

The handover should identify the source retained, imported/rejected counts, outstanding exceptions, responsible owner, recovery route and deletion/retention decisions. Do not leave live exports in a public repository or an engineer’s forgotten temporary directory.

Sources and review

Prepared by Yellow Brick Tech. General guidance; the source and test conditions for your own situation still matter.

Next review: 4 May 2027

Related services

How Yellow Brick Tech can help.

Next step

Need to apply this to a real situation?

Ask a question or choose the closest project route. A short conversation can often clarify the next step.