`emmeans` interface fails when visit variable is contained only in interaction term

```r library(mmrm) fit <- mmrm( FEV1 ~ ARMCD + FEV1_BL:AVISIT + us(AVISIT | USUBJID), data = fev_data ) summary(fit) library(emmeans) emmeans(fit, ~ARMCD) # Errors: # Error in `beta_hat[kept] <- component(object, "beta_est")`: # ! NAs are not allowed in subscripted assignments # But it should work, see e.g. lm(): fit2 <- lm(FEV1 ~ ARMCD + FEV1_BL:AVISIT, data = fev_data) emmeans(fit2, ~ARMCD) # If we just omit that interaction term it works: fit3 <- mmrm( FEV1 ~ ARMCD + us(AVISIT | USUBJID), data = fev_data ) emmeans(fit3, ~ARMCD) # If we add a main effect of AVISIT it also works: fit4 <- mmrm( FEV1 ~ ARMCD + AVISIT + FEV1_BL:AVISIT + us(AVISIT | USUBJID), data = fev_data ) emmeans(fit4, ~ARMCD) ``` Session info: ``` R version 4.5.2 (2025-10-31) Platform: aarch64-apple-darwin20 Running under: macOS Tahoe 26.4.1 Matrix products: default BLAS: /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib LAPACK: /Library/Frameworks/R.framework/Versions/4.5-arm64/Resources/lib/libRlapack.dylib; LAPACK version 3.12.1 locale: [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8 time zone: Asia/Taipei tzcode source: internal attached base packages: [1] stats graphics grDevices utils datasets methods base other attached packages: [1] emmeans_2.0.3 mmrm_0.3.17 loaded via a namespace (and not attached): [1] vctrs_0.7.2 nlme_3.1-168 cli_3.6.5 TH.data_1.1-5 rlang_1.2.0 estimability_1.5.1 processx_3.8.6 [8] generics_0.1.4 xtable_1.8-8 glue_1.8.0 zoo_1.8-15 backports_1.5.1 TMB_1.9.19 ps_1.9.1 [15] pak_0.9.2 grid_4.5.2 MASS_7.3-65 mvtnorm_1.3-6 lifecycle_1.0.5 compiler_4.5.2 multcomp_1.4-30 [22] codetools_0.2-20 sandwich_3.1-1 coda_0.19-4.1 Rcpp_1.1.1 lattice_0.22-7 R6_2.6.1 splines_4.5.2 [29] pillar_1.11.1 Rdpack_2.6.6 rbibutils_2.4.1 callr_3.7.6 checkmate_2.3.4 Matrix_1.7-4 tools_4.5.2 [36] survival_3.8-3 ```