I was wondering why my mshapiro_test() call below throws an error?
I want this call to exclude variables ID and Group in my data.
library(tidyverse)
library(rstatix)
d <- read_csv("https://raw.githubusercontent.com/rnorouzian/v/main/memory.csv")
d %>% group_by(Group) %>% select(-ID) %>% mshapiro_test()
#Error: Problem with `mutate()` input `data`.
#x is.numeric(x) is not TRUE
We can use cur_data() within summarise
library(dplyr)
d %>%
select(-ID) %>%
group_by(Group) %>%
summarise(out = list(mshapiro_test(cur_data())))
-output
# A tibble: 3 x 2
# Group out
#* <chr> <list>
#1 CON <tibble [1 × 2]>
#2 HI <tibble [1 × 2]>
#3 MT <tibble [1 × 2]>
The out is a list column. It can be unnested
library(tidyr)
d %>%
select(-ID) %>%
group_by(Group) %>%
summarise(out = list(mshapiro_test(cur_data()))) %>%
unnest(c(out))
# A tibble: 3 x 3
# Group statistic p.value
# <chr> <dbl> <dbl>
#1 CON 0.907 0.122
#2 HI 0.963 0.599
#3 MT 0.947 0.412
The cur_data() only selects the columns that are not the grouping
d %>%
select(-ID) %>%
group_by(Group) %>%
summarise(out = list(head(cur_data(), 2))) %>%
pull(out)
[[1]]
# A tibble: 2 x 2
MP SE
<dbl> <dbl>
1 4 6
2 6 4
[[2]]
# A tibble: 2 x 2
MP SE
<dbl> <dbl>
1 10 7
2 7 8
[[3]]
# A tibble: 2 x 2
MP SE
<dbl> <dbl>
1 6 9
2 7 9
Or another option is group_split with map
library(purrr)
d %>%
select(-ID) %>%
group_split(Group, .keep = FALSE) %>%
map_dfr(mshapiro_test)
# A tibble: 3 x 2
# statistic p.value
# <dbl> <dbl>
#1 0.907 0.122
#2 0.963 0.599
#3 0.947 0.412
Or using base R with split and sapply/lapply
sapply(split(d[3:4], d$Group), mshapiro_test)
# CON HI MT
#statistic 0.9070134 0.9627111 0.9470829
#p.value 0.1218592 0.5993172 0.4120478