问题
I have a dataset with IDs and Value where one ID could take multiple values. Currently, the same ID is repeated row-wise when it has multiple values, but I hope to keep one ID per row, adding more columns when necessary. Here's a reproducible example:
df <- data.frame(id = c(1,1,1,2,3,3),
val = c(10:15))
What I want is
df2 <- data.frame(id = c(1:3),
val1 = c(10, 13, 14),
val2 = c(11, "", 15),
val3 = c(12, "", ""))
I've tried to get it done using pivot_wider as below but it had two problems: (1) couldn't figure out how to create a column name of my own; (2) wanted to add columns but it creates a list within a cell.
library(tidyr)
df %>% pivot_wider(values_from = val, values_fn = list, names_prefix = "x")
回答1:
You need to add a sequence column:
df %>% group_by(id) %>%
mutate(time=row_number()) %>%
pivot_wider(names_from=time, values_from = val, names_prefix = "val")
# A tibble: 3 x 4
# Groups: id [3]
id val1 val2 val3
<dbl> <int> <int> <int>
1 1 10 11 12
2 2 13 NA NA
3 3 14 15 NA
library(tidyr)
library(dplyr)
来源：https://stackoverflow.com/questions/62978811/r-pivot-wider-to-keep-one-id-per-row