如何在R语言中对同一Item下的多行数据进行筛选

我有一组被试作答数据，每个被试的每个Item对应3行记录，分别对应Item的3个不同Region。Observation字段表示被试是否注视该区域（0=未注视，1=注视）。需要筛选出满足以下所有条件的Region 3的RT数据：

• Region 1的Observation为1
• Region 2的Observation为0
• Region 3的Observation为1

比如示例数据里，Item 1的Region 3不满足条件（Observation为0），不会被纳入；Item 2的Region 3符合所有条件，会被保留。

之前尝试跨行筛选失败：能标记各区域的观测状态，但提取Region 3数据后无法关联同一Item下其他区域的状态，导致没有有效结果。试过按Item分组，但不知道后续操作。

示例数据 #

R数据结构 #

structure(list(Participant = c("p1", "p1", "p1", "p1", "p1", "p1"),
               Item = c("Item 1", "Item 1", "Item 1", "Item 2", "Item 2", "Item 2"),
               Region = c("Region 1", "Region 2", "Region 3", "Region 1", "Region 2", "Region 3"),
               Observation = c(1, 0, 0, 1, 0, 1),
               RT = c(3300L, NA, NA, 5000L, NA, 7000L)),
          class = "data.frame", row.names = c(NA, -6L))

解决方案 #

方法1：分组筛选（dplyr） #

按Participant + Item分组，在组内验证三个区域的观测条件，直接筛选符合要求的Region 3记录：

library(dplyr)

# 加载数据
df <- structure(list(Participant = c("p1", "p1", "p1", "p1", "p1", "p1"),
                     Item = c("Item 1", "Item 1", "Item 1", "Item 2", "Item 2", "Item 2"),
                     Region = c("Region 1", "Region 2", "Region 3", "Region 1", "Region 2", "Region 3"),
                     Observation = c(1, 0, 0, 1, 0, 1),
                     RT = c(3300L, NA, NA, 5000L, NA, 7000L)),
                class = "data.frame", row.names = c(NA, -6L))

# 筛选符合条件的Region 3数据
filtered_df <- df %>%
  group_by(Participant, Item) %>%
  filter(
    any(Region == "Region 1" & Observation == 1),
    any(Region == "Region 2" & Observation == 0),
    Region == "Region 3" & Observation == 1
  ) %>%
  ungroup()

# 输出结果
print(filtered_df)

运行结果：

# A tibble: 1 × 5
  Participant Item    Region   Observation    RT
  <chr>       <chr>   <chr>          <dbl> <int>
1 p1          Item 2  Region 3          1  7000

方法2：宽表转换筛选（tidyr + dplyr） #

先将长表转为宽表，把每个Item的三个区域数据合并到一行，再筛选条件，最后提取需要的Region 3数据：

library(tidyr)
library(dplyr)

df <- structure(list(Participant = c("p1", "p1", "p1", "p1", "p1", "p1"),
                     Item = c("Item 1", "Item 1", "Item 1", "Item 2", "Item 2", "Item 2"),
                     Region = c("Region 1", "Region 2", "Region 3", "Region 1", "Region 2", "Region 3"),
                     Observation = c(1, 0, 0, 1, 0, 1),
                     RT = c(3300L, NA, NA, 5000L, NA, 7000L)),
                class = "data.frame", row.names = c(NA, -6L))

# 转宽表
wide_df <- df %>%
  pivot_wider(
    id_cols = c(Participant, Item),
    names_from = Region,
    values_from = c(Observation, RT),
    names_sep = "_"
  )

# 筛选条件并提取数据
filtered_result <- wide_df %>%
  filter(
    Observation_Region 1 == 1,
    Observation_Region 2 == 0,
    Observation_Region 3 == 1
  ) %>%
  select(Participant, Item, RT_Region 3) %>%
  # 可选：转回长表格式
  rename(RT = RT_Region 3) %>%
  mutate(Region = "Region 3", Observation = 1) %>%
  select(Participant, Item, Region, Observation, RT)

print(filtered_result)

运行结果和方法1一致。

内容的提问来源于stack exchange，提问作者user20792955