R语言：按Species分组合并Events/Groups均有匹配的数据框行

核心实现思路为：按物种拆分后，以“两行同时存在Events公共元素、Groups公共元素”作为连接边，构建邻接图后识别连通分量，对每个连通分量内的元素完成去重、排序后逗号拼接，完全匹配需求规则。

可复现代码 #

首先载入你提供的原始数据，完成基础清洗（转字符串类型、去除分组字段的前置空格）：

# 原始数据导入
df <- structure(list(Species = structure(c(3L, 3L, 2L, 2L, 1L, 1L), .Label = c("Dog", 
"Elephant", "Monkey"), class = "factor"), Family = structure(1:6, .Label = c("A", 
"A,B", "B", "C", "K", "L,M,N"), class = "factor"), Events = structure(c(2L, 
3L, 4L, 6L, 1L, 5L), .Label = c("10", "6,7", "6,8,9", "7,8", 
"8,10,9", "9,10"), class = "factor"), Groups = structure(c(1L, 
3L, 2L, 4L, 5L, 6L), .Label = c(" G1,G2", " G6,G7", "G1,G2,G4,G8,G12", 
"G6", "G90", "G90,G91"), class = "factor")), class = "data.frame", row.names = c(NA, 
-6L))

# 基础数据清洗
df[] <- lapply(df, as.character)
df$Groups <- trimws(df$Groups) # 去除原始Groups字段的前置空格

载入依赖包，执行合并逻辑：

library(igraph)
library(dplyr)

final_res <- df %>%
  group_split(Species) %>%
  lapply(function(species_part) {
    # 单条数据直接返回
    if(nrow(species_part) == 1) return(species_part)

    # 拆分每行的Events、Groups为独立元素集合
    event_list <- strsplit(species_part$Events, ",", fixed = TRUE)
    group_list <- strsplit(species_part$Groups, ",", fixed = TRUE)
    row_num <- nrow(species_part)

    # 构建邻接关系：满足双交集条件的两行标记为连通
    adj_matrix <- matrix(FALSE, row_num, row_num)
    for (i in 1:(row_num-1)) {
      for (j in (i+1):row_num) {
        has_event_inter <- length(intersect(event_list[[i]], event_list[[j]])) > 0
        has_group_inter <- length(intersect(group_list[[i]], group_list[[j]])) > 0
        if (has_event_inter & has_group_inter) {
          adj_matrix[i,j] <- adj_matrix[j,i] <- TRUE
        }
      }
    }
    diag(adj_matrix) <- TRUE # 行自身与自身连通

    # 识别连通分量，作为合并分组的唯一标识
    conn_graph <- graph_from_adjacency_matrix(adj_matrix)
    merge_group_id <- components(conn_graph)$membership

    # 按连通组合并元素：去重、排序后逗号拼接
    species_part %>%
      mutate(merge_group = merge_group_id) %>%
      group_by(Species, merge_group) %>%
      summarise(
        Family = paste(sort(unique(unlist(strsplit(Family, ",", fixed=TRUE)))), collapse = ","),
        Events = paste(sort(as.integer(unique(unlist(strsplit(Events, ",", fixed=TRUE))))), collapse = ","),
        Groups = paste(sort(unique(unlist(strsplit(Groups, ",", fixed=TRUE)))), collapse = ","),
        .groups = "drop"
      ) %>%
      select(-merge_group)
  }) %>%
  bind_rows()

输出结果 #

执行上述代码后得到的结果与你描述的期望输出完全一致：

细节说明 #

• 采用连通分量算法而非简单两两合并，支持传递性关联的合并场景：比如行1与行2符合合并条件、行2与行3符合合并条件时，即使行1与行3无直接公共元素，也会被合并到同一组，避免逻辑漏洞
• Events列做了数值类型转换后排序，不会出现字符串排序导致的顺序错乱（比如不会把"8,9,10"错误拼接为"8,10,9"）
• 自动过滤原始数据中的无效空格、重复元素，保证拼接结果干净无冗余

内容的提问来源于stack exchange，提问作者chippycentra