使用 dplyrivot_longer 转置数据集

Question

我有一个数据集，其中列是不同的站点，前两行提供了纬度和经度值。我想转置这些数据，以便每个站点现在都是一行，列中包含纬度和经度值。

我正在尝试使用 执行此操作pivot_longer，但迄今为止尚未成功，因为从示例中我不清楚如何指示哪些字段应该是新的行和列。

df <- data.frame(
  sites = c("lat", "lon"),
  A = c(10, 20),
  B = c(12, 18),
  C = c(14, 17),
  D = c(21, 12),
  E = c(3, 23)) %>% 
  # transpose with sites in 1st column (A-E on different rows) and lat/lon values in seperate columns
  pivot_longer(cols = c(2:6),
               names_to = c("lat", "lon"), 
               values_to = "sites")

Error in `build_longer_spec()`:
! If you supply multiple names in `names_to` you must also supply one of `names_sep` or `names_pattern`.
Run `rlang::last_error()` to see where the error occurred.

Answer 1

transpose我们这里可能需要

\n

library(data.table)\ndata.table::transpose(setDT(df), make.names = "sites")\n

\n

-输出

\n

     lat   lon\n   <num> <num>\n1:    10    20\n2:    12    18\n3:    14    17\n4:    21    12\n5:     3    23\n

\n\n

如果我们需要列名作为每行的标识符

\n

 data.table::transpose(setDT(df), make.names = "sites", keep.names = "grp")\n      grp   lat   lon\n   <char> <num> <num>\n1:      A    10    20\n2:      B    12    18\n3:      C    14    17\n4:      D    21    12\n5:      E     3    23\n

\n\n

如果我们想使用tidyverse，则将形状重塑为“长”，然后重塑为“宽”pivot_wider

\n

library(dplyr)\nlibrary(tidyr)\ndf %>% \n  pivot_longer(cols = -sites, names_to = 'grp') %>% \n  pivot_wider(names_from = sites, values_from = value)\n# A tibble: 5 \xc3\x97 3\n  grp     lat   lon\n  <chr> <dbl> <dbl>\n1 A        10    20\n2 B        12    18\n3 C        14    17\n4 D        21    12\n5 E         3    23\n

\n