problem with missing values in non normally distributed vars

[e joffe]

Hi,

I am using compareGroups to analyze a dataset

I have several numeric variables which have missing values.

When I run CG I get an error for all the values which have NAs but only for non-normally distributed 

when I use na.action=na.exclude it removes all the records because each record has a missing value somewhere.

any idea how to solve this ?

CG=compareGroups(conditioning_prot~age_trans+sex+type+subtype+stage_diag+b_symp+cns+hctci+protocol_1+

                   rad+numCyc_1+outcome_1+relapse_1+protocol_2+numCyc_2+outcome_2+

                   status_trans+time_from_dx+time_from_rel+mobilization_type+               dose_modification+modification_percent+nucleated_cells+cd34_kg+cd34_lt2+anc_engraftment+plt_engraft_25+

                   plt_engraft_50+relapse_post_trans+status+cause_death+mucositis+bacteremia_fungemia+mech_vent+

                   blood_trans+plt_trans+hospital_los,method=method,Q1=0, Q3=1,data=tecam)

Thanks,

Erel

[Isaac Subirana]

Could you provide an example data set in order to investigate what is happening?

Thanks.

[e joffe]

Hi,

So I have a problem with getting descriptive stats for non-normally distributed values.

Attached is the file with the data.

The code I am using is 

library(compareGroups)

compareGroups(AF~age,method=2,data=temp,Q1=0,Q3=1)

The error I get is

Error in compareGroups.default(X = X, y = y, include.label = include.label,  : 

  None variable can be computed

Thanks 

Erel

[compareGroups]

Hi Erel,

I've just reproduced your example and obtained no error but only a warning. Bellow, there are the results from my PC (Windows) with the latest R version (3.2.2) and latest compareGroups version (3.0.1)

Please, could you indicate me the R and compareGroups package versions you are using?

Best,

Isaac.

> library(compareGroups)

> compareGroups(AF~age,method=2,data=temp,Q1=0,Q3=1)

-------- Summary of results by groups of 'AF'---------

  var N   p.value method                selection

1 age 701 0.004** continuous non-normal ALL      

-----

Signif. codes:  0 '**' 0.05 '*' 0.1 ' ' 1 

Warning message:

In cor.test.default(x, as.integer(y), method = "spearman") :

  Cannot compute exact p-value with ties