从data.table多列提取多数值并生成新列的技术需求

针对给定的data.table数据框，需要从a至e列中提取每行的多数值（仅取值0或4）并生成新列，以下是两种高效实现方法：

方法一：行求和判断法 #

通过统计每行中4的出现次数，若次数≥3（5列的多数阈值）则取4，否则取0：

library(data.table)

# 构建示例数据
df <- data.table(id = 1:7,
                 reg = c("DD","BB","MA","CA","MN","KA","LA"),
                 a = c(4,4,0,4,0,4,0),
                 b = c(0,4,0,4,0,4,0),
                 c = c(4,0,4,4,4,4,0),
                 d = c(0,4,0,4,0,4,4),
                 e = c(0,4,4,4,0,4,0),
                 card = c("two","three","one","one","two","three","six")
)

# 添加多数值列
df[, majority := ifelse(rowSums(.SD == 4) >= 3, 4, 0), .SDcols = a:e]

方法二：逐行计算众数法 #

利用apply结合table统计每行的众数，适合扩展到多取值场景：

df[, majority := apply(.SD, 1, function(x) {
  count_table <- table(x)
  as.integer(names(count_table)[which.max(count_table)])
}), .SDcols = a:e]

最终输出结果 #

执行上述任意方法后，数据框会新增majority列，结果如下：

id reg a b c d e   card majority
1:  1  DD 4 0 4 0 0    two        0
2:  2  BB 4 4 0 4 4  three        4
3:  3  MA 0 0 4 0 4    one        0
4:  4  CA 4 4 4 4 4    one        4
5:  5  MN 0 0 4 0 0    two        0
6:  6  KA 4 4 4 4 4  three        4
7:  7  LA 0 0 0 4 0    six        0

内容的提问来源于stack exchange，提问作者samrr_tr