ols fails in 4.4-1, while working in 4.3-1

[RomanTsegelskyi]

It seems that in 4.4-1 ols fails for the same code that works in 4.3-1

v.4.4-1

> sessionInfo()
...
other attached packages:
[1] pander_0.6.0    rms_4.4-1       SparseM_1.6     Hmisc_3.17-0    ggplot2_1.0.1   Formula_1.2-1  
[7] survival_2.38-3 lattice_0.20-33

> set.seed(123)
> n <- 1000
> age <- rnorm(n, 50, 10)
> cholesterol <- rnorm(n, 200, 25)
> sex <- factor(sample(c("female", "male"), n, TRUE))
> health <- data.frame(age, cholesterol)
> dd <- datadist(age, sex) # Is needed for summary() below
> options(datadist = 'dd')    
> ols(cholesterol ~ age + sex)
subscript out of bounds
1: withCallingHandlers(eval(code, new_test_environment), error = capture_calls, message = function(c) invokeRestart("muffleMessage"))
2: eval(code, new_test_environment)
3: eval(expr, envir, enclos)
4: ols(cholesterol ~ age + sex)

v.4.3-1

> package_url <- 'https://cran.r-project.org/src/contrib/Archive/rms/rms_4.3-1.tar.gz'
> install.packages(package_url, repos=NULL, type="source")
...
> sessionInfo()
...
other attached packages:
[1] pander_0.6.0    rms_4.3-1       SparseM_1.6     Hmisc_3.17-0    ggplot2_1.0.1   Formula_1.2-1  
[7] survival_2.38-3 lattice_0.20-33

> set.seed(123)
> n <- 1000
> age <- rnorm(n, 50, 10)
> cholesterol <- rnorm(n, 200, 25)
> sex <- factor(sample(c("female", "male"), n, TRUE))
> health <- data.frame(age, cholesterol)
> dd <- datadist(age, sex) # Is needed for summary() below
> options(datadist = 'dd')    
> ols(cholesterol ~ age + sex)
Linear Regression Model

ols(formula = cholesterol ~ age + sex)

                 Model Likelihood     Discrimination    
                    Ratio Test           Indexes        
Obs      1000    LR chi2      7.73    R2       0.008    
sigma 25.1697    d.f.            2    R2 adj   0.006    
d.f.      997    Pr(> chi2) 0.0210    g        2.501    

Residuals

      Min        1Q    Median        3Q       Max 
-75.32269 -17.42416  -0.05717  17.86732  82.67025 

          Coef     S.E.   t     Pr(>|t|)
Intercept 190.4515 4.2067 45.27 <0.0001 
age         0.2188 0.0803  2.72 0.0066  
sex=male   -0.7503 1.5932 -0.47 0.6378  

[harrelfe]

This issue should have been reported under the rms project. The problem is that I did not update the version of Hmisc required by rms in the rms DESCRIPTION file. If you update Hmisc to the version that is on CRAN this problem goes away.

[RomanTsegelskyi]

Sorry for filing the issue at a wrong place and thanks for a quick answer