Testing a mapping with real data

The built-in tester and sample payloads — and choosing the samples that would actually catch something, which is the whole of this lesson. The sample everybody uses is the one they were given: a tidy record from the partner, one line, every field populated, produced by somebody demonstrating that the file format works. It passes against almost any mapping, correct or not. THE FIVE SAMPLES THAT EARN THEIR PLACE. ONE: NULLS. Every optional element absent. This is what proves Lesson 3's silent-default behaviour and shows you what the target fills in. TWO: EMPTIES. Elements present but blank — which is a different case from absent, and the two behave differently on an update. THREE: MAXIMUM LENGTHS. The longest name, the longest description, the longest reference. Truncation is silent in most mappers, and a target field two characters shorter than the source is a defect that only shows on real data.…

In this module: Module 5 · Mappings and transformations

1. 1What a mapping really isFree preview
2. 2The mapper, practically
3. 3Required, optional and the empty element
4. 4Repeating elements and loops
5. 5Conditional mapping
6. 6Functions: strings, numbers and dates
7. 7Dates and time zones, carefully
8. 8Numbers, precision and currency
9. 9Lookups inside a mapping
10. 10Testing a mapping with real data
11. 11Namespaces, and the errors they cause
12. 12Keeping a mapping maintainable
13. 13What breaks: five mapping failures
14. 14Lab briefing · Map two schemas that disagree