DOB → Age conversion

Any column whose header (case-insensitively, after trimming) equals dob is replaced in the output with a column called Age.

Algorithm

1. The cell value is passed to new Date(...). Anything Date can parse — ISO strings, US- and EU-style formatted strings, Excel-numeric serials surfaced as strings — is accepted.
2. If parsing yields Invalid Date, the row's Age cell becomes empty ("") and the per-sheet invalidDobCount increments.
3. Otherwise, age is computed in whole years from "today" (server clock), with a real birthday-not-yet-passed-this-year check that subtracts one year if the birthday hasn't happened yet.
4. The result is sanity-clamped to 0 ≤ age < 130. Anything outside that range is treated as an invalid DOB and replaced with "".

The implementation is in calculateAgeFromDOB — it's the source of truth.

What you get

Input cell	Output Age
1990-01-15	35 (or 34, depending on today's date relative to Jan 15)
15/01/1990	Same — the Date constructor handles common formats.
not-a-date	"" and invalidDobCount += 1
empty	"" (no penalty — empty cells are not "invalid")
1820-01-01	"" — age would exceed 130, treated as invalid.

Why "Age" and not "DOB"

DOB is a quasi-identifier — combined with postcode and gender it's enough to re-identify a sizeable fraction of a population. Age in whole years is far less revealing while preserving the clinical signal most analyses actually need.

Caveats

• The age is computed at upload time using the server's local clock. The same workbook re-uploaded in a later year will yield older ages. If you need a stable snapshot age, persist the manifest and the CSVs together — the manifest's generatedAt timestamp tells you when the ages were computed.
• There is no support for fractional ages (months) or for clinical "decimal age" conventions used in pediatric work. Add those as a follow-up rule if needed.