【发布时间】：2020-10-17 15:13:23
【问题描述】：
我正在处理潜水行为，然后我有一个函数可以在数据框中创建一个新列，其中包含每次潜水的太阳角度的表盘相位（白天、黎明、夜晚和黄昏），但显然它们的条件是重叠的彼此。我的真实数据框有超过 90.000 行
(1) 日出时，黎明时太阳高程为 -12 至 +6 度，相对于地平线，(2) 黄昏时太阳高程为 +6 至 -12 度，日落时。 (3) 大于+6度时确定白天，(4)小于-12度时确定夜晚。
dt = data.table(localtime= c("2016-10-24 12:45:06", "2016-10-24 12:46:13", "2016-10-24 12:47:02", "2016-10-24 12:48:27", "2016-10-24 12:52:39", "2016-10-24 12:55:11", "2016-10-30 21:08:02", "2016-10-30 21:18:27", "2016-10-30 21:30:13","2016-10-24 23:27:21", "2016-10-26 06:54:29"),
lon = c(-39.94400, -39.94410, -39.94418, -39.94432, -39.94472, -39.94496, -40.87705, -40.87567, -40.87409, -40.00234, -40.59864),
DepthMean = 30, 50, 200, 76, 467, 87, 98, 10, 240, 176, 89))
dialphase <- function(df) {
#loading the necessary package
require(oce)
#getting the sun angle from local time, longitude and latitude by the correction for atmospheric refraction
x2<-sunAngle(df$localtime, df$lon, df$lon, useRefraction = TRUE)
#converting into data.frame
x3<-do.call(rbind, lapply(x2$altitude, as.data.frame))
#creating new colum with the sun angles
df$sun<- x3$`X[[i]]`
#creating new column with only by hour about local time to do the next conditions
df$hourBRT <- as.POSIXlt(df$localtime)$hour
df <- df[!is.na(df$hourBRT),]
#creating new column with dial phases by sun angles and hours with day, dawn, night and dusk
df$dial_phase[df$sun >= 6.0] <- "day"
df$dial_phase[df$sun < 6.0 & df$sun > -12.0 & df$hourBRT > 3] <- "dawn" #manhã
df$dial_phase[df$sun <= -12.0] <- "night"
df$dial_phase[df$sun < 6.0 & df$sun > -12.0 & df$hourBRT > 16 ] <- "dusk" #noite
df$hourBRT<-NULL
return(df)
}
我预料到了：
dt
localtime lon DepthMean dial_phase
1: 2016-10-24 12:45:06 -39.94400 30 day
2: 2016-10-24 12:46:13 -39.94410 50 day
3: 2016-10-24 12:47:02 -39.94418 200 day
4: 2016-10-24 12:48:27 -39.94432 76 day
5: 2016-10-24 12:52:39 -39.94472 467 day
6: 2016-10-24 12:55:11 -39.94496 87 day
7: 2016-10-30 21:08:02 -40.87705 98 dusk
8: 2016-10-30 21:18:27 -40.87567 10 dusk
9: 2016-10-30 21:30:13 -40.87409 240 dusk
10: 2016-10-24 23:27:21 -40.00234 176 night
11: 2016-10-26 06:54:29 -40.59864 89 dawn
>
我的猜测是因为使用if、else 或ifelse 时该功能会正常工作，我对吗？我没有使用这些代码执行功能的能力。有人可以帮忙吗？
我试过这个东西，但是缺少一个阶段
> f1<-function(df) {
+ #loading the necessary package
+ require(oce)
+
+ #getting the sun angle from local time, longitude and latitude by the correction for atmospheric refraction
+ x2<-sunAngle(df$localtime, df$lon, df$lon, useRefraction = TRUE)
+
+ #converting into data.frame
+ x3<-do.call(rbind, lapply(x2$altitude, as.data.frame))
+
+ #creating new colum with the sun angles
+ df$sun<- x3$`X[[i]]`
+
+ #creating new column with only by hour about local time to do the next conditions
+ df$hourBRT <- as.POSIXlt(df$localtime)$hour
+ df <- df[!is.na(df$hourBRT),]
+ df$dial_phase <- with(df, ifelse(sun >= 6.0, "day",
+ ifelse(sun < 6.0 & sun > -12.0 & hourBRT > 3, "dawn",
+ ifelse(sun <= -12.0, "night",
+ ifelse(sun < 6.0 & sun > -12.0 & hourBRT > 16 , "dusk", NA_character_)))))
+ df
+ }
> g<-f1(df)
> table(df$dial_phase) ### dusk is missing
dawn day night
12185 46276 33593
** 编辑 **
我猜我的功能不行，因为当我绘图时，表盘相位重叠，像这样：
df$hourBRT <- as.POSIXlt(df$localtime)$hour
df <- df[!is.na(df$hourBRT),]
ggplot(df, aes(as.factor(hourBRT), DepthMean, col = dial_phase)) +
geom_boxplot() +
scale_y_log10(breaks = c(10, 50, 100, 200, 300, 400, 500, 600)) +
xlab("Hour of the day") +
ylab("Depth of Dives (m)")
【问题讨论】：
-
请展示一个可重现的小例子
-
我试试，等一下
-
在最后一部分中，您没有更新
with(df, ...并且在此之前返回了原始 'df' -
我在下面发布了一个解决方案。由于没有数据，我无法测试它，但我猜它的方向是正确的
-
我编辑了帖子，请看看df是否有效
标签： r function if-statement conditional-statements