group_by is not working as it expected

[sai_matcha]

Hello,

I have a dataframe look like this

df <-  data.frame(TIME = rep(1:1:24,3),
                  DV = runif(72, min=5, max=99))

i want to get the statistical summary of DV at each time point
i have used the below code, which worked fine earlier , but now it is not grouping but gives the over all result

df2 = df %>%
  group_by(TIME)%>%
  summarise(q5 = quantile(DV,0.05),
            q50 = quantile(DV,0.5),
            q95 = quantile(DV,0.95))

print(df2)
        q5      q50      q95
1 12.27739 53.75661 91.69368

can someone helps me to get summary at each time point .

Thnaks in advance

[EconProf]

I think your code works fine for me:

library(tidyverse)

df <-  data.frame(TIME = rep(1:1:24,3),     # why not just rep(1:24, 3)?
                  DV = runif(72, min=5, max=99))

df %>%
  group_by(TIME)%>%
  summarise(q5 = quantile(DV,0.05),
            q50 = quantile(DV,0.5),
            q95 = quantile(DV,0.95))
#> # A tibble: 24 × 4
#>     TIME    q5   q50   q95
#>    <int> <dbl> <dbl> <dbl>
#>  1     1  29.0  73.6  82.1
#>  2     2  68.2  74.0  94.3
#>  3     3  77.6  87.5  91.2
#>  4     4  18.5  77.2  77.9
#>  5     5  14.0  39.3  61.9
#>  6     6  39.2  83.1  91.1
#>  7     7  15.4  95.6  96.6
#>  8     8  29.3  57.3  63.0
#>  9     9  31.3  60.7  61.4
#> 10    10  22.5  60.7  84.7
#> # ℹ 14 more rows

^{Created on 2023-12-13 with reprex v2.0.2}

[jrkrideau]

Change the name of your data.frame. df is a function name and you may have some kind of a conflict. Otherwise I agree with EconProf. Your code is working fine for me.

[sai_matcha]

Hi, @EconProf @jrkrideau thanks for your help. I have figured something. Actually the code works fine, but it is problem of loading multiple packages. Code is not working when i load both "dplyr" and "plyr" packages, but it works if i load only "dplyr" package.

Thanks again for your help!

[jrkrideau]

Ah, that sounds reasonable. {dplyr} is descended from {plyr}. There likely are conflicting names.

[EconProf]

Did you reverse the order the two packages are loaded? Try it just to satisfy my curiosity!

[jrkrideau]

Sounds likely. I load {data.table} before {tidyverse} in order to not mess up {lubridate}. Humm, maybe I should learn more about {data.table}'s time and date functions.

[jrkrideau]

Oh BTW here is a simple {data.table } equivalent:

library(data.table)
dat1  <-  data.frame(TIME = rep(1:1:24,3), DV = runif(72, min=5, max=99))
DT  <- as.data.table(dat1)

DT[ ,  .(q5 = quantile(DV,0.05),
             q50 = quantile(DV,0.5),
            q95 = quantile(DV,0.95)),
           by = TIME]

[nirgrahamuk]

There is a package called conflicted that can be useful in these sorts of issues

conflicted: An Alternative Conflict Resolution Strategy