Error when summarise refers to previously created variable

[hadley]

library(dtplyr)
library(dplyr, warn.conflicts = FALSE)

lz <- lazy_dt(data.frame(x = 1:10))
lz %>% summarise(x = mean(x), y = x + 1) %>% collect()
#>       x  y
#>  1: 5.5  2
#>  2: 5.5  3
#>  3: 5.5  4
#>  4: 5.5  5
#>  5: 5.5  6
#>  6: 5.5  7
#>  7: 5.5  8
#>  8: 5.5  9
#>  9: 5.5 10
#> 10: 5.5 11

^{Created on 2019-06-26 by the reprex package (v0.2.1.9000)}

Use same technique from dbplyr.

[dyrland]

I know this is closed, but I use summarise() all the time in the wild.

I often need to express variables in different ways, as different bosses like data presented differently. This leads to me doing multiple summaries for the same variables. For instance, below I need the number of Auto Lunches per Manager and then the percent that are Auto Lunches. I prefer the first, df.out, (it's the cleanest, IMHO) but the second, dt.out, is ok. The third is also fine for this MWE, but as the calculations become more complex repeating the code becomes unintelligible.

Of course, perhaps I could have data prep skills. :D

df.out <- df %>%
  group_by(Manager) %>%
  summarize(`Auto Lunches` = sum(auto.lunch == TRUE),
            `Gross Hours` = sum(auto.lunch.time) / -60,
            `Auto Lunch %` = `Auto Lunches` / n() * 100,
            `Auto Lunch % (Mods)` = `Auto Lunches` / sum(modified == TRUE) * 100
  )
   
dt <- lazy_dt(df)
dt.out1 <- dt %>%
  group_by(Manager) %>%
  summarize(`Auto Lunches` = sum(auto.lunch == TRUE),
            `Gross Hours` = sum(auto.lunch.time) / -60,
            n = n()) %>% 
  mutate(`Auto Lunch %` = `Auto Lunches` / n * 100,
         `Auto Lunch % (Mods)` = `Auto Lunches` / sum(modified == TRUE) * 100
  )

dt.out2 <- dt %>%
  group_by(Manager) %>%
  summarize(`Auto Lunches` = sum(auto.lunch == TRUE),
            `Gross Hours` = sum(auto.lunch.time) / -60,
            `Auto Lunch %` = sum(auto.lunch == TRUE) / n() * 100,
            `Auto Lunch % (Mods)` = sum(auto.lunch == TRUE) / 
              sum(modified == TRUE) * 100
  )

[dyrland]

Yikes! dt.out1 won't work because I have to create a modified variable. I guess another reason why I am hopeful that I can use previously created variables in summarise()...

[gpierard]

I had this error because my df was a data.table. Got it fixed after converting to tibble

[TheDohn]

Could anyone elaborate on what:

Use same technique from dbplyr.

means? I can see how to avoid the error all together by calculating each summarize variable independently as it's own table, then joining the results together, but I am wondering if there is a better way using dtplyr.