Mood's pairwiseMedianTest in R - how get a nice output in Alteryx?

[enylen]

Hi!

As I am quite new here I would like to get all the help I can get. Sorry if this is the wrong thread. This is about R and Alteryxx

I try to compare medians and as Alteryx from what I know do not have a function for that, I tried to import some R libraries (RVAideMemoire and rcompanion) to try to use the functions Mood medtest and Pairwise Medtest to do so. I have figured out how to get them to run, but I have not figured out how I could make a neat output to some tables. For Mood's medtest it does not matter much since I will only get one R-value, but from the Pairwise Medtest I would like to get the output in order like something like this for the Pairwise Medtest:

Comparison	p.value
A - B	0.41...
A - C	0.73...
A - D	0.73..
B - C	0.31...
B - D	1
C - D	1

That is very similar to what the program prints in the "Message screen" (see below)

R (1) Comparison p.value p.adjust

R (1) 1 A - B = 0 0.4142 1

R (1) 2 A - C = 0 0.7389 1

R (1) 3 A - D = 0 0.7389 1

R (1) 4 B - C = 0 0.3173 1

R (1) 5 B - D = 0 1 1

R (1) 6 C - D = 0 1 1

Since I am a novice however, I only manage to get the p.values printed out in this fashion (see below), which severely impacts the ability to use the function/R-program for 50+ different examination.

X.0.4142.	X.0.7389.	X.0.7389.1	X.0.3173.	X.1.	X.1..1
0.4142	0.7389	0.7389	0.3173	1	1

Do you have any ideas on how to get to the first table?

Best regards,

an R noob in need

[JohnJPS]

Hi @enylen,tried running it and couldn't get the packages imported into Alteryx; (our security is something else).

However, I did get it running in R-Studio, and I don't see the amount of detail shown in your question...

Did you have any additional code to support what referenced as showing in your "Message screen" and/or the output you managed to get?

[enylen · Accepted solution]

Hi!

This is what I did to solve it: 

1. Import the packages (https://community.alteryx.com/t5/Alteryx-Designer-Knowledge-Base/Install-R-Packages/ta-p/41265)
2. Run the code below and use the function capture to get all output data albeit in a very unstructured way 
3. Use functions in Alteryx to clean the data
   1. Transpose
   2. Regex_Replace Regex_Replace([Value],"\s{1,}",'|')
   3. Column Split to divide it up by |
   4. Select interesting comments

In order to make it work I had to keep the names in the column type without spaces. Otherwise the Alteryx output would not be able to be handled by Regex_replace above. I also kept the names very short. For some strange reason long names would split the output that belonged together (name of pairwise comparison, p value, adjusted p value) into different rows even if they belonged to the same comparison.   

Code used:
input.data <- read.Alteryx.First("#1", 5000, mode = "data.frame")

if(!require(rcompanion)){install.packages("rcompanion")}
library(rcompanion)

PT = pairwiseMedianTest(Value ~ Type,
     data = input.data,
     exact = NULL,
     method = "fdr")
     # Adjusts p-values for multiple comparisons;
    # See ?p.adjust for options

PT #not sure if necessary
ut <- capture.output(PT)

data.out1 <- ut

write.Alteryx(data.out1, nOutput = 1)