使用$运算符

时间:2017-07-04 05:47:03

标签: r operators

我的问题是关于访问我的数据帧特定部分的$运算符。

以下是我的数据框

structure(list(Treatment = structure(c(11L, 11L, 11L, 11L, 11L, 
11L), .Label = c("Chlora", "Gen", "Lin", "Metro", "Metro+Pen", 
"Metro+Rif", "Metro+Rif+Pen", "Pen", "Pen+Rif", "Rif", "Untreated"
), class = "factor"), CDW = c(0.833, 0.867, 0.833, 2.7, 3.3, 
3.133), Chlorophyll = c(1.098, 1.376, 1.64, 34.714, 37.773, 40.54
), EPS = c(0.092, 0.067, 0.092, 0.571, 0.591, 0.597), Day = c(0L, 
0L, 0L, 10L, 10L, 10L), TreatmentDay = structure(c(21L, 21L, 
21L, 22L, 22L, 22L), .Label = c("Chlora0", "Chlora10", "Gen0", 
"Gen10", "Lin0", "Lin10", "Metro+Pen0", "Metro+Pen10", "Metro+Rif+Pen0", 
"Metro+Rif+Pen10", "Metro+Rif0", "Metro+Rif10", "Metro0", "Metro10", 
"Pen+Rif0", "Pen+Rif10", "Pen0", "Pen10", "Rif0", "Rif10", "Untreated0", 
"Untreated10"), class = "factor"), richness = c(36L, 37L, 39L, 
31L, 31L, 24L), shannon = c(3.35867443756547, 3.37354007038303, 
3.43640985397324, 3.16018680691703, 3.15718684762791, 2.85288614815093
)), .Names = c("Treatment", "CDW", "Chlorophyll", "EPS", "Day", 
"TreatmentDay", "richness", "shannon"), row.names = c("Day0F1", 
"Day0F2", "Day0F3", "Day10F1", "Day10F2", "Day10F3"), class = "data.frame")

我可以用

Unt$richness
[1] 36 37 39 31 31 24

但是,如果我想进行anova测试,我会收到一条错误消息,说我不能使用原子矢量$

Unt_Richness_anova <- aov(Unt$richness, TreatmentDay)

所以,我觉得很困惑我可以在一个实例中使用原子向量中的$运算符但不能用于另一个实例。 有人可以帮忙吗?

1 个答案:

答案 0 :(得分:0)

通过查看aov()函数的源代码,它实际上是对另一个名为terms()的函数的调用,导致此错误。

现在,如果在terms()函数中未指定数据框时调用aov()函数,则terms()函数似乎只会采用公式aov()函数传递的样式参数。

请注意,以下代码不起作用(如您所指定):

Unt_Richness_anova <- aov(Unt$richness, Unt$TreatmentDay)

如果将两个变量表示为公式,那么一切都按预期工作:

Unt_Richness_anova <- aov(Unt$richness ~ Unt$TreatmentDay)

问题不在于$运算符,而在于terms()函数。 terms()函数似乎只能处理单独的变量,如果它们属于同一数据框 AND ,则在aov()函数中指定数据框。

为什么会出现这种情况我无法回答,但希望其他人可以解释一下?