Data & Database Testing Services | ETL, Migration, BI QA Skip to main content
QAonTOP
Contact Us

Data and database testing, from the first row to the final report

Data and database testing checks that data is stored correctly, moves between systems without loss, stays accurate and secure, and shows the right numbers in every report. We test schemas, queries, pipelines, migrations and dashboards, then report every issue with the query and the rows that prove it.

  • 12 data platform types
  • 20 testing types
  • 22 databases and data platforms
  • Free one-week POC

[ Data platform types ]

Twelve kinds of data platform, each with its own failure points

An order database fails when two people save at once. A warehouse fails when a join drops rows. A dashboard fails quietly, with a number that looks right. We start with the risks that belong to your platform, and test those first.

01

Transactional apps (OLTP)

App databases behind orders, bookings, accounts and payments.

Top 3 risks we test first

  • Transactions that save halfway
  • Two users updating the same record
  • Business rules enforced in the database
02

Data warehouses

Central stores of history used for reporting.

Top 3 risks we test first

  • Facts and dimensions joining correctly
  • Loads finishing on time
  • History kept as designed
03

Data lakes / lakehouses

Raw and curated data in open files and tables.

Top 3 risks we test first

  • Raw, cleaned and curated layers agreeing
  • Unreadable or mis-partitioned files
  • Who can read which zone
04

ETL / ELT pipelines

Jobs that move and reshape data between systems.

Top 3 risks we test first

  • Rows lost or doubled in transit
  • Transform rules applied wrongly
  • Failed runs that don’t alert
05

Streaming / real-time data

Events and changes processed as they happen.

Top 3 risks we test first

  • Lost, repeated or out-of-order events
  • Lag during spikes
  • Producers changing the event format
06

BI dashboards & reports

Charts and KPIs leaders make decisions on.

Top 3 risks we test first

  • Totals that don’t match the source
  • Filters and date ranges miscounting
  • Users seeing figures they shouldn’t
07

Database migrations & upgrades

Moving to a new engine, version or schema.

Top 3 risks we test first

  • Changed behaviour after the upgrade
  • Slower queries on the new version
  • A rollback that doesn’t work
08

Legacy → cloud data migration

Old on-premise data moving to a cloud database.

Top 3 risks we test first

  • Every record and relationship arriving
  • Hidden rules in the old system
  • Downtime longer than the window
09

Master data / CRM & ERP data

Customer, product and supplier records shared across systems.

Top 3 risks we test first

  • Duplicate customers and products
  • Systems disagreeing on the same record
  • Sync jobs overwriting good data
10

NoSQL apps

Apps built on document, key-value or wide-column stores.

Top 3 risks we test first

  • Documents missing required fields
  • Replicas returning different answers
  • NoSQL injection through queries
11

AI / ML data pipelines

Data prepared to train and feed models.

Top 3 risks we test first

  • Bad rows reaching training data
  • Features differing between training and live
  • Personal data in training sets
12

Multi-tenant SaaS databases

One database shared by many customers.

Top 3 risks we test first

  • One tenant’s rows visible to another
  • Schema changes across every tenant
  • One heavy tenant slowing the rest

[ What we test ]

Twenty testing types, sorted by the layer they protect

Data problems rarely start where they show up. We group our testing into five layers, from how data is shaped to who can reach it, so you can see exactly what each test protects.

Layer 01

Structure

How data is shaped

Schema & structure testing

Tables, columns, data types, keys and indexes checked against the design, and schema changes checked before they ship.

Data integrity & constraint testing

Primary and foreign keys, unique and check constraints, and defaults tested so bad data can’t get in.

Stored procedure, trigger & function testing

Logic inside the database tested with known inputs and expected outputs, including errors and edge cases.

CRUD & business rule testing

Create, read, update and delete actions from the app checked in the database, along with the business rules behind them.

More on this service →
Layer 02

Movement

How data travels

ETL / ELT pipeline testing

Extract, transform and load jobs tested step by step: mappings, joins, lookups, type conversions and failed runs.

Source-to-target reconciliation

Row counts, totals and sample records compared between source and target, so nothing is lost, doubled or changed.

Data migration testing

Every record, relationship and piece of history checked after a move, plus a tested rollback.

Streaming & CDC testing

Change data capture and event streams checked for lost, repeated and out-of-order events, and for lag.

Layer 03

Quality

What people read and trust

Data quality testing

Completeness, accuracy, consistency, validity, uniqueness and timeliness measured against agreed thresholds.

BI & report testing

Dashboard and report figures checked against hand-written queries, including filters, dates and drill-downs.

Layer 04

Resilience

How it holds up

Database performance & query tuning

Slow queries found and measured, query plans and indexes reviewed, and improvements re-measured.

More on this service →

Load & concurrency testing

Many users and jobs reading and writing at once, checked for locks, deadlocks, timeouts and lost updates.

More on this service →

Backup, restore & recovery testing

Backups restored for real and recovery time and data loss measured against your targets.

Replication & failover testing

Replicas checked for lag and consistency, and failover tried to see what happens to live traffic.

Upgrade & version compatibility testing

Database engine and driver upgrades tested for changed behaviour, deprecated features and performance shifts.

Across all layersRegression automation for dataSQL and pipeline checks that run automatically on every change or scheduled load, so regressions surface the same day. Checks live in your repository and run in your CI or scheduler.More on test automation →

[ Databases & data quality ]

“Correct” data, measured six ways

Data can load without errors and still be wrong. We agree a threshold for each of six quality dimensions with you, check every one automatically and report a score, with the rows that fail.

  1. 01

    Completeness

    No missing rows or empty required fields.

  2. 02

    Accuracy

    Values match the real-world source.

  3. 03

    Consistency

    The same fact agrees across every system.

  4. 04

    Validity

    Formats, ranges and codes follow the rules.

  5. 05

    Uniqueness

    No duplicate customers, orders or products.

  6. 06

    Timeliness

    Data arrives when people need it.

Databases and data platforms we test

Relational
PostgreSQLMySQLMariaDBMicrosoft SQL ServerOracleSQLiteIBM Db2Amazon Aurora / RDSAzure SQLGoogle Cloud SQL
NoSQL & search
MongoDBRedisCassandraDynamoDBElasticsearch
Warehouses & lakehouses
SnowflakeBigQueryAmazon RedshiftDatabricks
Pipelines & streaming
Apache KafkaApache Airflowdbt

How we keep your data safe while testing

  • NDA signed before we get any access
  • Masked or synthetic copies of your data by default
  • Only the access the testing needs, removed when we finish
  • Read-only checks on production only when you agree
  • Every query, script and report belongs to you

[ Process & deliverables ]

Five steps, and what you get at each one

You always know what we’re doing and what you’ll get. Every query, script and report belongs to you.

  1. Step 1 of 5

    Discover

    We map your sources, pipelines, stores and reports, and the business rules your data must follow.

    You getTest strategy & data map
  2. Step 2 of 5

    Plan

    We agree mapping and reconciliation rules, quality thresholds and a safe test data approach.

    You getData test plan & mapping sheetTest data approach
  3. Step 3 of 5

    Test

    SQL checks, pipeline runs and report comparisons, with every issue logged as it’s found.

    You getBug reports with query and sample rowsDaily or weekly status reports
  4. Step 4 of 5

    Specialist checks

    Performance, concurrency, security and recovery testing where your scope includes them.

    You getPerformance reportSecurity reportRecovery test report
  5. Step 5 of 5

    Automate & sign off

    Data checks wired into your CI or scheduler, fixes retested and the release or migration signed off.

    You getAutomated data checksReconciliation & sign-off report
  • Free one-week POCSee how we test your data before you commit.
  • Your tools, your workflowWe work in your Jira, repository, CI and scheduler.
  • Evidence, not opinionsEvery bug comes with the query and the rows that show it.

[ Tools ]

The tools behind our data and database testing

Most data testing is careful SQL. We add tools where they make checks faster, repeatable or automatic, and we’re happy to use the ones you already have.

SQL & scripting

03 TOOLS
  • SQL
  • Python
  • pandas

Database unit testing

03 TOOLS
  • tSQLt
  • pgTAP
  • utPLSQL

Data quality & pipeline checks

03 TOOLS
  • Great Expectations
  • Soda Core
  • dbt tests

Pipelines & streaming

03 TOOLS
  • Apache Airflow
  • dbt
  • Apache Kafka

Performance

03 TOOLS
  • JMeter
  • HammerDB
  • pgbench

Security

03 TOOLS
  • sqlmap
  • OWASP ZAP
  • Burp Suite

Database clients

04 TOOLS
  • DBeaver
  • pgAdmin
  • SQL Server Management Studio
  • MongoDB Compass

Test data

02 TOOLS
  • Faker
  • Masked database copies

CI/CD & test management

06 TOOLS
  • Jenkins
  • GitHub Actions
  • GitLab CI
  • Jira
  • TestRail
  • Xray

See the full list on our technology stack page →

[ Case Studies ]

Data we’ve tested

Products where the data had to be right, written up by the engineers who tested them.

“QAonTOP has provided at least 20 times the value. Their testing service is outstanding and has greatly streamlined our process.”
Techup Software SolutionDirector and CEO, Software Service Provider

All case studies →

[ Ways to work ]

Six ways to get your data and databases tested

Choose the engagement model that matches how your data changes. You can switch models as your platform grows.

Not sure which fits? Find your model in four questions or compare all six.

[ FAQ ]

Data and database testing questions, answered

{{ f.a }}

[ Contact us ]

Get a test plan for your data

✓ Free one-week POC✓ NDA before access✓ A QA lead replies the same day

[ Keep exploring ]

Other platforms we test

All platforms →

Also useful: ERP & CRM testing · All testing services · Fixed-price packages · Estimate your QA cost