consider adding option to include missing report for all variables

[ericpgreen]

The nhanes_2010 dataset has 1417 obs, but the summary table will indicate a total of 940 obs (complete cases) unless we specify na.rm=FALSE. The documentation says this about the na.rm parameter:

when set to FALSE it also shows how many missing values are in the data for each categorical variable being summarized

And that's what this does...

library(furniture)
library(tidyverse)
data("nhanes_2010")

nhanes_2010 %>%
  furniture::table1("Age Mean (SD)" = age,                
         "Health" = gen_health,
         "Sex" = gender,
         "Cancer" = cancer, 
         "Asthma" = asthma,
         test = TRUE,                            
         output = "html",
         na.rm = FALSE,
         total = TRUE,
         type = "condense")                      

We see that 155 obs are missing on gen_health. But I'm wondering if there could be an option to also show missing for all variables. This table gives me the impression that we're only missing data for the gen_health variable, but that's not the case.

table(nhanes_2010$cancer, useNA = "always") returns 344 missing

table(nhanes_2010$asthma, useNA = "always") returns 2 missing

There are no missing on the numeric variable age, but that would be interesting to know too.

[TysonStanley]

Yes, this is something I've been wanting to work on but haven't had the time yet. Your post will help push it toward the top of the priority list. For a short-term fix (bc it would be more straightforward with formatting), I was thinking of adding the missing for the continuous variables next to the name of the variable. Something like:

----------------------------------
                      Mean/SD
Var1 (missing = 34)   2.5 (3.2)
----------------------------------

What do you think?

[ericpgreen]

Thanks! That seems reasonable to me. It would make for a really long label if Var1 is long, but certainly good in the short term.

On Mon, Jan 6, 2020 at 1:31 PM Tyson Barrett notifications@github.com wrote:

Yes, this is something I've been wanting to work on but haven't had the time yet. Your post will help push it toward the top of the priority list. For a short-term fix (bc it would be more straightforward with formatting), I was thinking of adding the missing for the continuous variables next to the name of the variable. Something like:

---

                  Mean/SD

Var1 (missing = 34) 2.5 (3.2)

What do you think?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/TysonStanley/furniture/issues/23?email_source=notifications&email_token=AAEFNAIX2R6ROK33TBSIRKDQ4N2G5A5CNFSM4KDGF6S2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEIGKOSY#issuecomment-571254603, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAEFNAKNHFRQKO2UPDQVKZLQ4N2G5ANCNFSM4KDGF6SQ .