Question

我需要对“ session_start”行之间的列journey中的步骤进行编号。我想不出如何为这种情况编写循环。

df <- data.table(
  page = c("page_1", "page_2", "page_3", "page_1", "page_2", "page_1", "page_2", "page_3"),
  journey = c("session_start", NA, NA, "session_start", NA, "session_start", NA, NA)      
)

所需的结果应该是这样。

df <- data.table(
  page = c("page_1", "page_2", "page_3", "page_1", "page_2", "page_1", "page_2", "page_3"),
  journey = c("session_start", "step_1", "step_2", "session_start", "step_1", "session_start", 
              "step_1", "step_2")

)

Answer 1

这就是您想要的。请确保将stringsAsFactors = F添加到data.table，否则journey列将无法正确处理。

for (i in 1:nrow(df)) {
    if (is.na(df$journey[i])) {
        df$journey[i] <- paste('step',step_index,sep='')
        step_index <- step_index + 1
    } else {
        step_index <- 1
    }
}

Answer 2

您可以尝试以下方法：

df$journey <- ifelse(df$page == "page_1","session_start", gsub(".*_","step_",df$page))

哪个给：

    > df
     page       journey
1: page_1 session_start
2: page_2        step_2
3: page_3        step_3
4: page_1 session_start
5: page_2        step_2
6: page_1 session_start
7: page_2        step_2
8: page_3        step_3

Answer 3

尝试使用ave的此解决方案。

i <- df$journey == "session_start"
i[is.na(i)] <- 0L
f <- cumsum(i)
df$journey <- ave(as.character(df$journey), f, FUN = function(s){
  s[is.na(s)] <- paste0("step_", seq_along(s)[-length(s)])
  s
})

使用data.table软件包可能有更好的方法，但是我不太熟练。

循环计算R中参数之间的行

3 个答案: