Collection
Diagnose data joins and quality before trusting a report
Turn quiet row multiplication, missing values, profiling limits and misleading quality metrics into testable decisions and repair ownership.
Explore in any order.
What’s included
3 checklists · 3 heuristics · 5 principles · 5 protocols
16 ideas
Predict rows from the join condition
A matching key does not promise a matching row.
Read the idea PrincipleKeep left rows when filtering
An invoice filter can make a supplier disappear.
Read the idea ProtocolAggregate each detail table before joining two totals
A join can inflate both totals at once.
Read the idea ProtocolDetect changes involving NULL
'Not equal' can be unknown, not true.
Read the idea ChecklistDefine what complete means
96% of what?
Read the idea PrincipleDo not mistake a valid format for a true value
Well-formed wrong data is still wrong.
Read the idea ProtocolTrace missing values
One percentage can hide where values disappear.
Read the idea HeuristicPrioritise errors by their impact
The busiest quality dashboard is not necessarily the safest system.
Read the idea PrincipleCheck records and fields separately
Filled fields cannot account for missing records.
Read the idea ChecklistCheck what Power Query profiles
A clean preview can leave rows unchecked.
Read the idea PrincipleRead distinct and unique in Power Query
Two similar words answer different counting questions.
Read the idea PrincipleCount rows and non-null values separately
A row can exist even when the invoice does not.
Read the idea HeuristicInspect failed data checks
The small failed slice may be the part that matters.
Read the idea ProtocolGive a data defect an owner
A logged error still needs a next step.
Read the idea ChecklistPublish the limits with the data extract
A dataset without its caveats invites incorrect decisions.
Read the idea ProtocolReplay examples before changing a data rule
Stricter can still be wrong.
Read the idea