如何简化R语言统计多列组合非空值及指定字符数的矩阵生成流程

实现思路 #

• 借助outer函数批量处理矩阵维度的两两计算，完全替代嵌套for循环
• 用列表存储A-H的统计结果，避免动态生成变量的冗余操作，调用更便捷

完整代码 #

# 示例数据（可直接运行）
example<-matrix(NA,40,7)
colnames(example)=c("month1","month2","month3","month4","month5","month6","month7")
example[,1]<-rep(c("A","B","C","D","E","F","G","H",NA),len=40)
example[,2]<-rep(c("A","H",NA,"C","E",NA,"F",NA),len=40)
example[,3]<-rep(c("C","A","C",NA),len=40)
example[,4]<-rep(c(NA,"A","F","H","E",NA),len=40)
example[,5]<-rep(c("A",NA),len=40)
example[,6]<-rep(c("C",NA,NA,"E"),len=40)
example[,7]<-rep(c("C","A",NA,NA,"C",NA,"F",NA),len=40)
example<-as.data.frame(example)
sub<-c("ALL","A","B","C","D","E","F","G","H")

# ----------------------
# 1. 生成ALLtable矩阵
# ----------------------
n_col <- ncol(example)
# 定义两两列共同非空行数统计函数
count_all <- Vectorize(function(i,j) {
  sum(!is.na(example[,i]) & !is.na(example[,j]))
})
ALLtable <- outer(1:n_col, 1:n_col, count_all)
# 给矩阵加行列名
dimnames(ALLtable) <- list(colnames(example), colnames(example))

# ----------------------
# 2. 生成A-H对应的Xtable矩阵
# ----------------------
sub_char <- sub[sub != "ALL"]
res_list <- lapply(sub_char, function(x) {
  count_x <- Vectorize(function(i,j) {
    # 统计第i列等于X时，i、j列共同非空的行数
    sum(example[,i] == x & !is.na(example[,j]), na.rm = TRUE)
  })
  xtab <- outer(1:n_col, 1:n_col, count_x)
  dimnames(xtab) <- list(colnames(example), colnames(example))
  return(xtab)
})
# 给列表命名，可直接通过res_list[["A"]]调取A对应的统计矩阵
names(res_list) <- sub_char

# 可选：如果需要单独生成Atable、Btable...全局变量，运行以下代码即可
# list2env(setNames(res_list, paste0(names(res_list), "table")), envir = .GlobalEnv)

结果说明 #

• ALLtable为7×7矩阵，行列名对应月度列名，ALLtable[i,j]即为第i和第j个月度列的共同非空行数
• res_list为命名列表，res_list[["A"]]到res_list[["H"]]分别对应A到H的统计矩阵，符合需求2的计算规则

内容的提问来源于stack exchange，提问作者D.PARK