added arguments for relabeling the "Missing" (NA) rows of table output.

[d-morrison]

This is an attempt to add an option to replace the label "Missing" for NA counts with a user-supplied string. It isn't quite working yet; I'm still trying to figure out the code structure. I thought I'd start the PR in case you might be willing to guide me.

[benjaminrich]

This is already possible using render.missing. Here is an example:

require(table1)
require(survival)

set.seed(789)

dat <- subset(survival::pbc, !is.na(trt))  # Exclude subjects not randomized

dat$trt     <- factor(dat$trt, levels=1:2, labels=c("D-penicillamine", "Placebo"))
dat$sex     <- factor(dat$sex, levels=c("m", "f"), labels=c("Male", "Female"))

# Add some missing values
dat$age[sample(1:nrow(dat), 11)] <- NA
dat$sex[sample(1:nrow(dat), 3)] <- NA

table1(~ age + sex | trt, data=dat, render.missing=c("XYZ"="FREQ (PCT%)"))

[d-morrison]

Thanks for the reply! I tried that code, and I see that it does use the custom label "XYZ"; however, it doesn't look like the parsing of "FREQ (PCT%)" is working:

#>                          D-penicillamine           Placebo           Overall
#> 1                                (N=158)           (N=154)           (N=312)
#> 2                  age                                                      
#> 3            Mean (SD)       51.2 (11.1)       48.4 (9.99)       49.8 (10.6)
#> 4    Median [Min, Max] 51.2 [26.3, 78.4] 47.6 [30.6, 74.5] 49.6 [26.3, 78.4]
#> 5                                                                           
#> 6                  XYZ       FREQ (PCT%)       FREQ (PCT%)       FREQ (PCT%)
#> 7                  sex                                                      
#> 8                 Male        21 (13.3%)         14 (9.1%)        35 (11.2%)
#> 9               Female       136 (86.1%)       138 (89.6%)       274 (87.8%)
#> 10                   X                                                      
#> 11              Male.1        21 (13.3%)         14 (9.1%)        35 (11.2%)
#> 12            Female.1       136 (86.1%)       138 (89.6%)       274 (87.8%)

I'll look through the documentation to see if I can figure out the needed format, but would be grateful for hints!

[benjaminrich]

Hmm... that is interesting, because the same code works for me, and produces this:

[d-morrison]

Installed version 1.4.3 from Github (upgrading from 1.4.2 from CRAN) and it works! Many thanks.