使用R中的data.table识别并注释组明确的最小值

Question

我需要确定样本数据中每个最小的value (忽略NA),class并在新列中将其标记为'min',如下所示data.table

样本数据:

df = structure(list(class = c("apple", "apple", "apple", "banana", 
"banana", "berry", "berry", "grape", "grape", "grape", "grape", 
"grape", "melon", "melon", "melon"), value = c(108816872, 108851837, 
108890411, 108784778, NA, 108784778, 108816872, 108816872, 108850460, 
NA, NA, NA, NA, NA, NA)), .Names = c("class", "value"), class = "data.frame", row.names = c(NA, 
-15L))

期望的输出:

#    class     value anno
#1   apple 108816872  min
#2   apple 108851837   NA
#3   apple 108890411   NA
#4  banana 108784778  min
#5  banana        NA   NA
#6   berry 108784778  min
#7   berry 108816872   NA
#8   grape 108816872  min
#9   grape 108850460   NA
#10  grape        NA   NA
#11  grape        NA   NA
#12  grape        NA   NA
#13  melon        NA   NA
#14  melon        NA   NA
#15  melon        NA   NA

Answer 1

我打算建议@eddies方法,但这是另一种选择

setDT(df)[order(value), min := c("min", rep(NA, .N - 1)), by = class]

编辑,如果你想要实际值而不是"min",你可以修改为

setDT(df)[order(value), min := c(value[1L], rep(NA, .N - 1L)), by = class]

Answer 2

dt = as.data.table(df) # or convert in place using setDT

dt[dt[, .I[which.min(value)], by = class]$V1, anno := 'min']