如何用dplyr/tidyr转换数据框：重构含metric列的表格

Got it, reshaping your wide data into the long format you need is straightforward using tidyr::pivot_longer() (part of the tidyverse, which works seamlessly with dplyr). Here's how to do it:

First, assuming you already have your dat data frame set up as you showed, you can use this code to pivot the numeric columns into a metric column while keeping your identifier columns (group, temp, temp2) intact:

library(tidyr)
library(dplyr)

# Convert wide to long format
dat_long <- dat %>%
  pivot_longer(
    cols = c(num, calc1, calc2),  # Which columns to pivot into the metric column
    names_to = "metric",          # Name of the new column for metric labels
    values_to = "value"           # Name of the new column holding the metric values
  )

Resulting Data Frame #

When you run this, dat_long will look like this:

# A tibble: 9 × 5
  group temp  temp2 metric value
  <chr> <chr> <chr> <chr>  <dbl>
1 A     GG    ll    num        1
2 A     GG    ll    calc1      4
3 A     GG    ll    calc2      7
4 B     HH    pp    num        2
5 B     HH    pp    calc1      5
6 B     HH    pp    calc2      8
7 C     KK    rr    num        3
8 C     KK    rr    calc1      6
9 C     KK    rr    calc2      9

Alternative (Older Function) #

If you're more familiar with the older gather() function (now superseded by pivot_longer()), you can also use this equivalent code:

dat_long <- dat %>%
  gather(key = "metric", value = "value", num, calc1, calc2)

Either way, you'll get the exact structure you wanted—with a metric column containing num, calc1, calc2, and all your original identifier columns preserved.

内容的提问来源于stack exchange，提问作者user3022875