如何仅因为另一个变量具有 NA 而从计算中排除所有行

Question

问题描述：我想应用一个公式，排除包含至少一个 NA 值的所有行。对于可重现的示例，我使用了以下数据框：

df <- data.frame(
           stringsAsFactors = FALSE,
                check.names = FALSE,
                       Date = c("01/11/1876",
                                "01/12/1876","01/01/1877","01/02/1877",
                                "01/03/1877","01/04/1877","01/05/1877",
                                "01/06/1877","01/07/1877","01/08/1877","01/09/1877",
                                "01/10/1877","01/11/1877","01/12/1877",
                                "01/01/1878","01/02/1878"),
        `Bissen P [mm]` = c(59.2,139.4,
                                89.6,46.9,41.1,105.1,45.3,62.3,NA,119.1,
                                121.2,86.1,42.1,54,37.3,71.8),
        `Bissen Q [mm]` = c(46.5,126.5,
                                111.7,53.9,49.4,28.1,31.9,22.3,19.1,24.9,
                                74.1,74.1,51.8,47.3,40.8,53.4),
    `Ettel P [mm]` = c(54,131.3,
                                83.3,42.1,NA,98.3,40.6,56.9,44.7,111.8,
                                113.8,79.9,37.5,49,32.8,66.1),
      `Ettel Q [mm]` = c(29.9,84.1,
                                77.9,39.9,35.1,21.7,24.4,15.7,14.5,20.1,
                                51.1,58.6,36,41.8,32.7,37.7),
           `Aalbach P [mm]` = c(46.6,112.7,
                                71.7,36.5,31.7,84.4,35.2,49.1,38.7,96,
                                97.7,68.7,32.6,42.3,28.5,57),
           `Aalbach Q [mm]` = c(59.8,132.4,
                                87.3,48.7,43.4,101.4,47.2,62.6,51.1,114.1,
                                116,84.1,44.3,55.1,39.9,71.2))

#Where I have 3 sites (Bissen, Ettel and Aalbach) and each site has 2 variables (P and Q)
#And I would like to obtain the runoff coefficient of each site (Q/P), 

在黄色矩形中，我已指示每个站点必须排除哪些行，因为一行包含 NA。

 
#step 2: extract site names
site_names <- sub(" P \\[mm\\]| Q \\[mm\\]", "", names(df)[-1]) |>
  unique()
site_names

#step 3: define function based on formula: 
#Here I have indicated that NA should be excluded, however, it does not eliminate all rows as I need. 
formula <- function(P, Q) {
  runoffcoefficient = mean(Q, na.rm = TRUE)/mean(P, na.rm = TRUE)
  return(data.frame(runoffcoefficient = runoffcoefficient))
}

# step 4: Loop through all sites
results <- list()
for (site in site_names) {
  site_data <- df[, grepl(site, names(df))]
  coefficients <- formula(site_data[[paste0(site, " P [mm]")]], 
                         site_data[[paste0(site, " Q [mm]")]]) 
  df$Site = site # Add the Site column to each site's anomalies
  results[[site]] <- coefficients
}  

> results  
$Bissen
  runoffcoefficient
1         0.7160308

$Ettel
  runoffcoefficient
1         0.5588475

$Aalbach
  runoffcoefficient
1          1.246611

如果消除了行，Bissen 将等于 0.7467，而 Ettel = 0.5624

任何帮助将不胜感激

Answer 1

你在追寻

complete.cases

df[complete.cases(df), ]