summarise() 使用 dplyr 的案例子集

Question

數據：

structure(list(subjnum = c(1L, 1L, 1L, 1L, 1L, 1L), expVers = structure(c(2L, 
2L, 2L, 2L, 2L, 2L), .Label = c("Angry", "Happy"), class = "factor"), 
    intendedSOA = c(1000L, 1000L, 100L, 100L, 50L, 50L), compatability = structure(c(1L, 
    1L, 1L, 1L, 1L, 1L), .Label = c("Comp", "Incomp"), class = "factor"), 
    T1RT = c(229L, 229L, 277L, 277L, 280L, 280L), T2RT = c(791L, 
    791L, 563L, 563L, 760L, 760L), T1ACC = c(1L, 1L, 1L, 1L, 
    1L, 1L), T2ACC = c(1L, 1L, 1L, 1L, 1L, 1L)), row.names = c(NA, 
6L), class = "data.frame")

我想通過它們的mean()來總結列 T1RT、T2RT、T1ACC 和 T2ACC，並通過數據中的其他變量/因素（subjnum、intendedSOA、兼容性、expVers）來組織這些值。 但是，變量 T1RT 和 T2RT 摘要的摘要不應包括T1ACC or T2ACC == 0的試驗，但變量 T1ACC 和 T2ACC 的摘要應包括所有值（無條件求和）。 我嘗試通過執行以下操作在summarise()中包含一個if()參數：

> backcomplong2 <- ACC %>%
+   select(subjnum, expVers,intendedSOA, compatability, T1RT, T2RT, T1ACC, T2ACC)%>%
+   group_by(subjnum, compatability, expVers, intendedSOA)%>%
+   summarise(T1RT = if(T1ACC == 1 && T2ACC == 1) round(mean(T1RT)), 
+             T2RT = if(T1ACC == 1 && T2ACC == 1) round(mean(T2RT)),
+             T1ACC = mean(T1ACC), 
+             T2ACC = mean(T2ACC))

但收到此錯誤：

Problem with `summarise()` input `T1RT`.
x Input `T1RT` must be a vector, not NULL.
i Input `T1RT` is `if (T1ACC == 1 && T2ACC == 1) round(mean(T1RT))`.
i The error occured in group 14: subjnum = 3, compatability = "Comp", expVers = "Happy", intendedSOA = 100.

****** 請注意，我的可重現數據不會返回相同的錯誤 ********
較大的數據（我沒有在此處提供，因為它太大而無法使用dput()粘貼到此問題中）返回錯誤。

我想我錯誤地使用了if()語句，也許我可以嘗試if_else()代替？ 另一種解決方法是簡單地執行summarise()兩次，一次用於 RT，另一次用於 ACC，但這更簡潔。

Answer 1

你放一個if來檢查一個條件，但不要放一個else並告訴當條件不滿足時需要做什么。 因此，它返回NULL object 導致錯誤。

如果您只對那些您想要取mean的值進行子集化，那么您不需要if / else在這里。 嘗試這個：

library(dplyr)

ACC %>%
   group_by(subjnum, compatability, expVers, intendedSOA)%>%
   summarise(T1RT = mean(T1RT[T1ACC == 1 & T2ACC == 1]), 
             T2RT =  mean(T2RT[T1ACC == 1 & T2ACC == 1]),
             T1ACC = mean(T1ACC), 
             T2ACC = mean(T2ACC))