[英]Using an ifelse function in a lapply/apply function
Trying to apply a function to a large dataset.尝试将 function 应用于大型数据集。 Specifically, trying to apply the mean of the lowest 1000 times (df$time) set before the date (df$date) found in that row.
具体来说,尝试应用在该行中找到的日期 (df$date) 之前设置的最低 1000 次 (df$time) 的平均值。 Applying this function on a small portion worked
将此 function 应用于一小部分工作
However, because the dataset is so large, I want to restrict the apply to just the 1% of rows where df$wr is true.但是,由于数据集太大,我想将应用限制在 df$wr 为 true 的 1% 的行中。
This is the code I wrote so far with mean1000 as the intended name of the new variable and the data set split based on name (25 categories):这是我到目前为止编写的代码,将 mean1000 作为新变量的预期名称,并根据名称拆分数据集(25 个类别):
df1 <- data.frame(
mean1000 = lapply(
split(df, df$name), function(y)
df$y$mean1000 = apply(y, 1, function(x) {ifelse(x["wr" == TRUE],
mean(sort(df$time[df$date < x["date"]])[2:1000]), NA)})) %>%
unlist()
)
Result:结果:
df1 is created, but it's just a table with 0 observations of 1 variable (mean1000) df1 已创建,但它只是一个表,其中包含 1 个变量 (mean1000) 的 0 个观察值
The error message is 25 times the following:错误消息是以下的 25 次:
1. Unknown or uninitialised column `y`.
I mostly followed the guidelines as outlined here , but those solutions are less complex/layered than what I'm trying to do.我主要遵循此处概述的指南,但这些解决方案没有我想要做的复杂/分层。 How can I adjust the code?
如何调整代码?
Data:数据:
| # | time | date | id1 | id2 | rank | name | wr |
|---|------|-----------|-----|-----|------|-------|------|
| 1 | 2408 | 2022-06-04| a8m2| pr9w| 24 | City01| TRUE |
| 2 | 2503 | 2022-06-25| b6p5| ur1r| 226 | City01| FALSE|
| 3 | 2672 | 2022-05-07| c8k1| py5l| 371 | City01| FALSE|
The desired result is to have an extra column added in which the mean calculated ( mean(sort(df$time[df$date < x["date"]])[2:1000]) ) is added when the wr value is TRUE.期望的结果是添加一个额外的列,当wr值为真的。
Consider by
(object-oriented wrapper to tapply
) which is very similar to split
+ lapply
but more streamlined.考虑
by
(object-oriented wrapper to tapply
),它与split
+ lapply
非常相似,但更精简。 Then run an embedded sapply
for rowwise mean conditional calculations.然后运行嵌入式
sapply
进行逐行平均条件计算。
# SORT DATA BY NAME AND DATE
df1 <- with(df1, df1[order(name, date),]) |> `row.names<-`(NULL)
# CONDITIONALLY CALCULATE MEAN BY GROUP
df1$mean100 <- by(df1, df1$name, function(sub), {
# ITERATE THROUGH EVERY DATE ROW
mean1000 <- sapply(
sub$date,
# SUBSET AND CALCULATE MEAN
FUN=\(dt) mean(sub$time[sub$date < dt][2:1000], na.rm=TRUE)
)
# CONDITIONALLY ADJUST BY wr FLAG
mean1000 <- ifelse(sub$wr == TRUE, mean1000, NA_real_)
})
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.