英文:
Get variable value in subset function - R
问题
在尝试获取子集函数中变量值时，我发现了一个问题。当运行代码时，我收到以下消息：“警告：Error in -: 无效的一元运算符参数”，因为子集函数“-c(val)”中的“val”未在上面定义为变量。cname <- c("A1","A2","A3","A4","A5","A6","A7","A8","A9","A10","A11","A12","A13","A14","A15","A16","A17","A18","A19","A20","A21","A22","A23","A24","A25","A26","A27","A28","A29","A30","A31")for (i in 15:length(cname)) {val <- cname[i]ifelse(sum(!is.na(df2$val))==0,df2 <- subset(df2, select = -c(val)),df2)}
df2 的结果是 此数据。
我的期望结果是删除仅包含 NA 值的不必要列，如您可以在 这里 看到的那样。
如何获取 val 的值，以便删除仅包含 NA 值的列？
<details><summary>英文:</summary>I found an issue while trying to get the value of a variable in the subset function. When I run the code, I receive the message: "Warning: Error in -: invalid argument to unary operator" because "val" in subset function "-c(val)" not define as variable above.
cname <- c("A1","A2","A3","A4","A5","A6","A7","A8","A9","A10",
"A11","A12","A13","A14","A15","A16","A17","A18","A19","A20",
"A21","A22","A23","A24","A25","A26","A27","A28","A29","A30","A31")
for (i in 15:length(cname)) {
val <- cname[i]
ifelse(sum(!is.na(df2$val))==0,
df2 <- subset(df2, select = -c(val)),
df2)
}
The df2 results in [this data][1].My expected result is to remove unnecessary columns that have NA values only, as you can see [here][2].How can I get the value from val, so I can remove the columns that have only NA values?[1]: https://i.stack.imgur.com/xOpmt.png[2]: https://i.stack.imgur.com/8pCkf.png</details># 答案1**得分**: 0We can use `subset` without a loop - use the vectorized `colSums` on a logical matrix (`is.na(df2)`) to return the count of NAs in each column, compare (`!=`) it with the number of rows (`nrow(df2)`) to create a logical vector, subset the column names, use that in `select` argument in `subset`:```Rsubset(df2, select = names(df2)[colSums(is.na(df2)) != nrow(df2)])
-output:
A1 A2 A4 A51 1 1 NA 102 2 2 NA 103 3 3 NA 104 4 NA 3 105 5 5 2 10
Or with tidyverse - use select and check for any non-NA elements in each column for selecting the column:
library(dplyr)df2 %>%select(where(~ any(!is.na(.x)))
-output:
A1 A2 A4 A51 1 1 NA 102 2 2 NA 103 3 3 NA 104 4 NA 3 105 5 5 2 10
data
df2 <- data.frame(A1 = 1:5, A2 = c(1:3, NA, 5), A3 = NA_integer_,A4 = c(NA, NA, NA, 3, 2), A5 = 10)
英文:
We can use subset without a loop - use the vectorized colSums on a logical matrix (is.na(df2)) to return the count of NAs in each column, compare (!=) it with the number of rows (nrow(df2)) to create a logical vector, subset the column names, use that in select argument in subset
subset(df2, select = names(df2)[colSums(is.na(df2)) != nrow(df2)])
-output
A1 A2 A4 A51 1 1 NA 102 2 2 NA 103 3 3 NA 104 4 NA 3 105 5 5 2 10
Or with tidyverse - use select and check for any non-NA elements in each column for selecting the column
library(dplyr)df2 %>%select(where(~ any(!is.na(.x))))
-output
A1 A2 A4 A51 1 1 NA 102 2 2 NA 103 3 3 NA 104 4 NA 3 105 5 5 2 10
data
df2 <- data.frame(A1 = 1:5, A2 = c(1:3, NA, 5), A3 = NA_integer_,A4 = c(NA, NA, NA, 3, 2), A5 = 10)</details>
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论