eduaguilera/whep
footprint_balance: compare_footprint_methods conflates NA-valued footprints with zero
オープン
#212 opened on 2026/07/22
area:footprintbuggood first issuehelp wantedmechanicalno-data-neededpriority:low
Repository metrics
- Stars
- (1 個のスター)
- PR merge metrics
- (PR metrics pending)
説明
What
compare_footprint_methods() (R/footprint_balance.R:104-121) uses sum(value) with no na.rm, so a cell containing an NA sums to NA; the subsequent tidyr::replace_na(value_a, 0) — intended to fill cells absent from one method — silently rewrites that NA to 0.
Why it matters
A genuine-but-NA footprint is reported as "zero footprint", fabricating a spurious abs_diff equal to the other method's full value.
Suggested fix / check
Distinguish "absent from this method" from "present but NA" before replacing with 0.
From deep static audit (2026-07-22), parallel subsystem review pass. Deduplicated against then-open issues. Not verified by running unless stated.