我对循环很陌生。我有一个难以处理的数据框架,我想要减少,这样只有观察(行)没有负数。我被困在这里了。这将在每次创建空值时创建一个空值,而不是裁剪后的数据帧。
mydata=for (i in names(df)) {
subset(df, df[[ paste(i)]]>=0)
}
发布于 2016-04-29 05:49:38
一个纯矢量化的解决方案怎么样:
DF[!rowSums(DF < 0), ]
# ID Items Sequence
#1 1 D 1
#2 1 A 2
#5 2 B 2
数据
DF=structure(list(ID = c(1, 1, 1, -1, 2), Items = c("D", "A", "A",
"A", "B"), Sequence = c(1, 2, -2, 1, 2)), .Names = c("ID", "Items",
"Sequence"), row.names = c(NA, -5L), class = "data.frame")
解释
比较DF < 0
为TRUE/FALSE
中的每个值提供data.frame
DF < 0
# ID Items Sequence
# [1,] FALSE FALSE FALSE
# [2,] FALSE FALSE FALSE
# [3,] FALSE FALSE TRUE
# [4,] TRUE FALSE FALSE
# [5,] FALSE FALSE FALSE
然后rowSums()
给出每一行的和(作为TRUE == 1, FALSE == 0
)
rowSums(DF<0)
# [1] 0 0 1 1 0
所以我们可以用这个向量来子集我们的data.frame。但是,因为我们希望它的值都是正的(即rowSums == 0),所以我们否定了过滤器。
DF[!rowSums(DF < 0), ]
发布于 2016-04-29 05:37:09
这不需要循环:)
DF=structure(list(ID = c(1, 1, 1, -1, 2), Items = c("D", "A", "A",
"A", "B"), Sequence = c(1, 2, -2, 1, 2)), .Names = c("ID", "Items",
"Sequence"), row.names = c(NA, -5L), class = "data.frame")
DF
# ID Items Sequence
#1 1 D 1
#2 1 A 2
#3 1 A -2
#4 -1 A 1
#5 2 B 2
new_DF = DF[apply(DF<0,1,function(x) !any(x)),]
new_DF
# ID Items Sequence
#1 1 D 1
#2 1 A 2
#5 2 B 2
https://stackoverflow.com/questions/36930138
复制相似问题