数据按时间顺序按条件排列(每个变量)，并根据条件在数据帧中循环(每个变量)

Question

我想按时间顺序将降雨时间序列数据组织到每个雨计(按代码列)中。雨量计在同一列中，与示例中的行相同，还有包含月份、年份和降雨信息的列。

在组织数据后，我需要执行统计测试，循环可以使它更容易，因为有大量的雨量计。是否有一种通过rain_gauge列中的代码循环指定每个雨量计作为测试中要重复的变量的方法？

rain_gauge = c(rep(1442032, 40), rep(1442035, 30), rep(1442036, 30),rep(1442039, 45),rep(1442049, 40),rep(1442032, 40),rep(1442045, 35))
year = runif(260, 1978,2020)
month = runif(260,1,12)
rainfall = runif(260, 50,202)

df = data.frame(rain_gauge, year, month, rainfall)#data frame to be organized in chronological order by "code" category
head(df)

#Examples of tests to apply to the series of each rain gauge in the rain_gauge column.
library(modifiedmk)
mmkh(as.vector(subset(df,rain_gauge=="1442032"))$rainfall)

pvalue_mk = mmkh(as.vector(subset(df,rain_gauge=="1442032"))$rainfall)[[2]]#Result to be save in a data frame results

library(tseries)
adf.test(subset(df,rain_gauge=="1442032")$rainfall)
pvalue_df = adf.test(subset(df,rain_gauge=="1442032")$rainfall)[[2]]#Result to be save in a data frame results

非常感谢！

Answer 1

考虑by (面向对象的tapply包装器)，它允许您按因子对数据帧进行切片，并在子集上运行进程以返回简化的对象(即向量、矩阵)或任何输出的列表：

library(modifiedmk)
library(tseries)

get_pvalues <- function(sub) {
    mmkh_obj <- mmkh(sub$rainfall)
    adf_obj <- adf.test(sub$rainfall)    

    # NAMED VECTOR
    c(pvalue_mmkh = mmkh_obj[, "p.value"], pvalue_adf = adf_obj[, "p.value"])
}

# NAMED MATRIX
pvalues_matrix <- by(df, df$rain_gauge, get_pvalues)