R 迭代筛选器列表

发布于12月14日

我想遍历应用于数据帧的多列的筛选器组合列表，如下所示:

df <- data.frame(    
    ID1 = c("Asset1", "Asset2", "Asset1", "Asset2", "Asset1", "Asset2", "Asset1", "Asset2"),
    ID2 = c("Asset1", "Asset1", "Asset2", "Asset2", "Asset1", "Asset1", "Asset2", "Asset2"),
    var1= c(0.011,      0.012,   0.012,    0.022,   .0011,      0.0012,   0.0012,    0.0022)
)

举个例子:

我想为以下组合计算var1的值:

filter1: ID1=='Asset1' 和 ID2=='Asset1'

和

filter2: ID1=='Asset1' 和 ID2=='Asset2'

对于一个组合，我显然可以这样做，用一个dspur过滤器:

library(dplyr)
df %>% filter(ID1=='Asset1', ID2=='Asset1')

但我想用一个遍历所有过滤器的for循环来完成这项工作:

filter_list <- c(filter1, filter2)

for (filter_i in filter_list){
    df %>% filter(filter_i)
}

推荐答案

您可以从列1:2中的unique个assets资源创建筛选器列表.接下来，使用expand.grid的所有2-排列生成filter_list.

> assets <- unique(unlist(df[1:2]))
> filter_list <- expand.grid(assets, assets) |> asplit(1)

最后，你可以lapply,subset以上的名单.

> lapply(filter_list, \(x) subset(x=df, ID1 == x[1] & ID2 == x[2]))
[[1]]
     ID1    ID2   var1
1 Asset1 Asset1 0.0110
5 Asset1 Asset1 0.0011

[[2]]
     ID1    ID2   var1
2 Asset2 Asset1 0.0120
6 Asset2 Asset1 0.0012

[[3]]
     ID1    ID2   var1
3 Asset1 Asset2 0.0120
7 Asset1 Asset2 0.0012

[[4]]
     ID1    ID2   var1
4 Asset2 Asset2 0.0220
8 Asset2 Asset2 0.0022

或者，如果你坚持for圈

> res <- vector('list', length(filter_list))  ## initialize
> for (i in seq_along(filter_list)) {
+   res[[i]] <- subset(x=df, ID1 == filter_list[[i]][1] & ID2 == filter_list[[i]][2])
+ }
> res
[[1]]
     ID1    ID2   var1
1 Asset1 Asset1 0.0110
5 Asset1 Asset1 0.0011

[[2]]
     ID1    ID2   var1
2 Asset2 Asset1 0.0120
6 Asset2 Asset1 0.0012

[[3]]
     ID1    ID2   var1
3 Asset1 Asset2 0.0120
7 Asset1 Asset2 0.0012

[[4]]
     ID1    ID2   var1
4 Asset2 Asset2 0.0220
8 Asset2 Asset2 0.0022

Data:个

> dput(df)
structure(list(ID1 = c("Asset1", "Asset2", "Asset1", "Asset2", 
"Asset1", "Asset2", "Asset1", "Asset2"), ID2 = c("Asset1", "Asset1", 
"Asset2", "Asset2", "Asset1", "Asset1", "Asset2", "Asset2"), 
    var1 = c(0.011, 0.012, 0.012, 0.022, 0.0011, 0.0012, 0.0012, 
    0.0022)), class = "data.frame", row.names = c(NA, -8L))