R中按组计算每两行值的比例

Question

我有这个数据集

\n

\n\ndf <- tibble(id, event, duration)\n

\n

我需要使用后续“表面”计算每个“潜水”行表面的持续时间比例，并将结果插入到新列中。所有这些都由“id”分隔。

\n

比例=水面/潜水+水面

\n

#Output dataframe\n\n# A tibble: 8 x 4\n  id    event   duration proportion    \n1 A     surface       56 x         \n2 A     surface       96 x         \n3 A     surface       14 x         \n4 A     surface       77 x         \n5 B     surface       28 x         \n6 B     surface       63 x         \n7 B     surface       47 x         \n8 B     surface       90 x   \n\n############################################################\n

\n

编辑：

\n

在我的原始数据中，我有一些没有“表面”的“潜水”，并且创建的此代码有错误。

\n

Error in `dplyr::mutate()`:\n! Problem while computing `proportion = DurationMin[What ==\n  "Surface"]/sum(DurationMin)`.\n\xe2\x9c\x96 `proportion` must be size 2 or 1, not 0.\n\xe2\x84\xb9 The error occurred in group 2803: ptt = "2017111870", grp = 1015.\n

\n

在“id”内将有奇数行，其中“潜水”事件在其序列中不会有“表面”。所以我需要每次遇到未配对的事件时，要么将其忽略，要么插入 NA。这是可能的？

\n

按照这个数据框示例：

\n

\nid <- c("A", "A", "A", "A", "A", "A", "A", "A", "B", "B", "B", "B", "B", "B", "B")\n\nevent <- c("dive", "surface", "dive", "surface", "dive", "surface", "dive", "surface", "dive", "surface", "dive", "surface", "dive", "surface", "dive")\n\nduration <- c(55, 56, 40, 96, 58, 14, 43, 77, 19, 28, 34, 63, 29, 47, 61)\n\ndf <- tibble(id, event, duration)\n\n> df\n   id   event duration\n1   A    dive       55\n2   A surface       56\n3   A    dive       40\n4   A surface       96\n5   A    dive       58\n6   A surface       14\n7   A    dive       43\n8   A surface       77\n9   B    dive       19\n10  B surface       28\n11  B    dive       34\n12  B surface       63\n13  B    dive       29\n14  B surface       47\n15  B    dive       61\n16  B    dive       45\n17  B surface       30\n> \n

\n

Answer 1

我们可以使用gl每 2 行创建分组索引，然后通过将事件值为 'surface' ( event == 'surface') 的 'duration' 除以sum'duration' 来创建列 'proportion'

\n

library(dplyr)\ndf %>%\n   group_by(id) %>%\n   group_by(grp = as.integer(gl(n(), 2, n())), .add = TRUE) %>% \n   mutate(proportion = duration[event == 'surface'][1]/sum(duration)) %>%\n   ungroup %>%\n   select(-grp)\n

\n

-输出

\n

# A tibble: 16 \xc3\x97 4\n   id    event   duration proportion\n   <chr> <chr>      <dbl>      <dbl>\n 1 A     dive          55      0.505\n 2 A     surface       56      0.505\n 3 A     dive          40      0.706\n 4 A     surface       96      0.706\n 5 A     dive          58      0.194\n 6 A     surface       14      0.194\n 7 A     dive          43      0.642\n 8 A     surface       77      0.642\n 9 B     dive          19      0.596\n10 B     surface       28      0.596\n11 B     dive          34      0.649\n12 B     surface       63      0.649\n13 B     dive          29      0.618\n14 B     surface       47      0.618\n15 B     dive          61      0.596\n16 B     surface       90      0.596\n

\n\n

对于新的数据集，我们可以使用

\n

df %>% \n  group_by(id) %>% \n  group_by(grp = cumsum(event == 'dive'), .add = TRUE) %>% \n  mutate(proportion = duration[event == 'surface'][1]/sum(duration)) %>% \n  ungroup %>%\n  select(-grp)\n

\n

-输出

\n

# A tibble: 17 \xc3\x97 4\n   id    event   duration proportion\n   <chr> <chr>      <int>      <dbl>\n 1 A     dive          55      0.505\n 2 A     surface       56      0.505\n 3 A     dive          40      0.706\n 4 A     surface       96      0.706\n 5 A     dive          58      0.194\n 6 A     surface       14      0.194\n 7 A     dive          43      0.642\n 8 A     surface       77      0.642\n 9 B     dive          19      0.596\n10 B     surface       28      0.596\n11 B     dive          34      0.649\n12 B     surface       63      0.649\n13 B     dive          29      0.618\n14 B     surface       47      0.618\n15 B     dive          61     NA    \n16 B     dive          45      0.4  \n17 B     surface       30      0.4  \n

\n

数据

\n

df <- structure(list(id = c("A", "A", "A", "A", "A", "A", "A", "A", \n"B", "B", "B", "B", "B", "B", "B", "B", "B"), event = c("dive", \n"surface", "dive", "surface", "dive", "surface", "dive", "surface", \n"dive", "surface", "dive", "surface", "dive", "surface", "dive", \n"dive", "surface"), duration = c(55L, 56L, 40L, 96L, 58L, 14L, \n43L, 77L, 19L, 28L, 34L, 63L, 29L, 47L, 61L, 45L, 30L)), \nclass = "data.frame", row.names = c("1", \n"2", "3", "4", "5", "6", "7", "8", "9", "10", "11", "12", "13", \n"14", "15", "16", "17"))\n

\n