重塑或 dcast 长到宽，没有 value.var 2 列

Question

我有一个 2 列的 data.frame df，前六行显示在这里，但它有更多的块序列，每个跨越 3 行：

blocksequenceid description
M049-S215-S085 ECDTM-49
M049-S215-S085 ICD-215 
M049-S215-S085 ICD-85
M049-S213-S044 ECDTM-49
M049-S213-S044 ICD-213
M049-S213-S044 ICD-44

我想把它转换成这种格式：

blocksequenceid description1 description2 description3 
M049-S215-S085 ECDTM-49 ICD-215 ICD-85
M049-S213-S044 ECDTM-49 ICD-213 ICD-44

我考虑过 dcast 和 reshape，但是当它显示 ERROR: column time not found for reshape 时我不知道该怎么做，而且我不确定 dcast 是否适合在这里使用。这是我尝试过的：

reshape(df, idvar='blocksequenceid', timevar = 'description', direction = 'wide')
reshape(df, idvar='blocksequenceid', v.names = 'description', direction = 'wide')

我确信这很简单，但我缺少一些东西。

Answer 1

这是可重现的数据。

t <- 'blocksequenceid description
M049-S215-S085 ECDTM-49
M049-S215-S085 ICD-215 
M049-S215-S085 ICD-85'

df <- read.table(text = t, header = T)

---

这是一个可能的解决方案。

library(tidyverse)
df %>% 
  rename(description1 = description) %>%
  mutate(description = row_number()) %>%
  spread(description, description1, sep = "")

#   blocksequenceid description1 description2 description3
# 1  M049-S215-S085     ECDTM-49      ICD-215       ICD-85

编辑修改后的数据

t <- 'blocksequenceid description
M049-S215-S085 ECDTM-49
M049-S215-S085 ICD-215 
M049-S215-S085 ICD-85
M049-S213-S044 ECDTM-49
M049-S213-S044 ICD-213
M049-S213-S044 ICD-44'

df <- read.table(text = t, header = T)

在更新的数据中，你应该先做group_by(blocksequenceid)。

library(tidyverse)
df %>% 
  rename(description1 = description) %>%
  group_by(blocksequenceid) %>%
  mutate(description = row_number()) %>%
  spread(description, description1, sep = "")

# # A tibble: 2 x 4
# # Groups:   blocksequenceid [2]
#   blocksequenceid description1 description2 description3
#   <chr>           <chr>        <chr>        <chr>       
# 1 M049-S213-S044  ECDTM-49     ICD-213      ICD-44      
# 2 M049-S215-S085  ECDTM-49     ICD-215      ICD-85