Re:Nobody ever got fired for using a struct
by generativenoise on Friday March 06, 2026 @12:04AM
(#47271104)
That is definitely waving the red flag of poor data design. It is doing the all I have is a single level hash table and no real arrays for your records. So you end up doing pseudo arrays like "Path Report Type 1", "Path Report Type 2", ...
To me this screams of being constrained to interop with spreadsheets, since that is when I have seen similar things.
Do you think working with the wide single level records is better/easier than a more decomposed design?
Also out of curiosity, what do they do when there are more than 5 Path Reports? Duplicate record or it just doesn't happen?
771 columns (and I've read the definitions for them all, plus about 50 more that have been retired). In the database, these are split across at least 3 tables (registry, patient, tumor). But when working with the records, it's common to use one joined table. Luckily, even that usually fits in RAM.