英文:
Combining column values with column names for some select columns using tidyr unite
问题
给定一个数据框：
df <- data.frame(Col1 = LETTERS[1:4], Col2 = LETTERS[23:26], Col3 = c(1:4), col4 = c(100:103))
我想要将列与它们的列名组合在一起。我知道可以使用tidyr中的unite函数，并获得以下输出：
df %>% unite(NewCol, c(Col1, Col4), remove = F)Col1 Col2 Col3 Col4 NewCol1 A W 1 100 A_1002 B X 2 101 B_1013 C Y 3 102 C_1024 D Z 4 103 D_103
但我想要将列名放在列的值旁边，如下所示（分隔符_实际上不是很重要）：
Col1 Col2 Col3 Col4 NewCol1 A W 1 100 Col1_A_Col4_1002 B X 2 101 Col1_B_Col4_1013 C Y 3 102 Col1_C_Col4_1024 D Z 4 103 Col1_D_Col4_103
我尝试了这里发布的解决方案，它确实产生了期望的输出，但创建了一个单独的输出。
imap_dfr(df %>% select(Col1, Col4), ~ paste(.y, .x, sep = "_")) %>%unite(NewCol, sep = "_")NewCol<chr>1 Col1_A_Col4_1002 Col1_B_Col4_1013 Col1_C_Col4_1024 Col1_D_Col4_103
您可以简单地使用bind_cols()将两者组合吗？如何确保两者之间保留了行的顺序？是否有另一种方法可以在同一个数据框中创建NewCol，类似于第一种情况中的unite？
您可以使用bind_cols()将两个数据框组合在一起，并确保它们的行顺序相同。以下是如何完成这个任务：
library(dplyr)library(tidyr)# 使用 unite 创建 NewColdf1 <- df %>%unite(NewCol, c(Col1, Col4), remove = FALSE)# 使用 imap_dfr 创建 NewColdf2 <- imap_dfr(df %>% select(Col1, Col4), ~ paste(.y, .x, sep = "_")) %>%rename(NewCol = .)# 使用 bind_cols 将两个数据框组合result_df <- bind_cols(df, df1["NewCol"], df2["NewCol"])# 打印结果print(result_df)
这将产生一个包含所需输出的数据框 result_df，并确保了行的顺序保持一致。
英文:
Given a dataframe:
df <- data.frame(Col1 = LETTERS[1:4], Col2 = LETTERS[23:26], Col3 = c(1:4), col4 = c(100:103))
I want to combine column with their column names. I know I can use unite from tidyr and get the following output.
df %>% unite(NewCol, c(Col1, Col4), remove = F)Col1 Col2 Col3 Col4 NewCol1 A W 1 100 A_1002 B X 2 101 B_1013 C Y 3 102 C_1024 D Z 4 103 D_103
But I want to have the column name next to the value of the column as follows (the separator _ is really not that important):
Col1 Col2 Col3 Col4 NewCol1 A W 1 100 Col1_A_Col4_1002 B X 2 101 Col1_B_Col4_1013 C Y 3 102 Col1_C_Col4_1024 D Z 4 103 Col1_D_Col4_103
I tried the solution posted here which does give the desired output but it creates a separate output.
imap_dfr(df %>% select(Col1, Col4), ~ paste(.y, .x, sep = "_")) %>%unite(NewCol, sep = "_")NewCol<chr>1 Col1_A_Col4_1002 Col1_B_Col4_1013 Col1_C_Col4_1024 Col1_D_Col4_103
Would I simply use bind_cols() to combine both? How do I know the sequence of the rows is preserved between the two? Is there another way that I can create NewCol within the same dataframe similar to unite in the first case?
答案1
得分: 2
一个选项是创建临时的'colname + value'列，然后在第二步中合并它们，例如：
## 加载库library(tidyverse)## 加载示例数据df <- data.frame(Col1 = LETTERS[1:4], Col2 = LETTERS[23:26], Col3 = c(1:4), Col4 = c(100:103))## 预期结果df %>%bind_cols(imap_dfr(df %>%select(Col1, Col4),~ paste(.y, .x, sep = "_")) %>%unite(newcol, sep = "_"))#> Col1 Col2 Col3 Col4 newcol#> 1 A W 1 100 Col1_A_Col4_100#> 2 B X 2 101 Col1_B_Col4_101#> 3 C Y 3 102 Col1_C_Col4_102#> 4 D Z 4 103 Col1_D_Col4_103## 对于少量列df %>%mutate(tmp_Col1 = paste0("Col1", "_", Col1),tmp_Col4 = paste0("Col4", "_", Col4)) %>%unite(newcol, c(tmp_Col1, tmp_Col4), sep = "_")#> Col1 Col2 Col3 Col4 newcol#> 1 A W 1 100 Col1_A_Col4_100#> 2 B X 2 101 Col1_B_Col4_101#> 3 C Y 3 102 Col1_C_Col4_102#> 4 D Z 4 103 Col1_D_Col4_103## 对于大量列df %>%mutate(across(c(Col1, Col4),~paste0(cur_column(), "_", .x))) %>%unite(newcol, c(Col1, Col4), sep = "_") %>%left_join(df)#> Joining with `by = join_by(Col2, Col3)`#> newcol Col2 Col3 Col1 Col4#> 1 Col1_A_Col4_100 W 1 A 100#> 2 Col1_B_Col4_101 X 2 B 101#> 3 Col1_C_Col4_102 Y 3 C 102#> 4 Col1_D_Col4_103 Z 4 D 103
创建于2023年06月22日，使用 reprex v2.0.2
如果你有大量要转换的列，使用across()可以让你使用tidyselect函数，比如starts_with()，来选择感兴趣的列，而不必逐个指定每列的名称。
英文:
One option is to create temporary 'colname + value' columns, then unite them in a second step, e.g.
## Load librarieslibrary(tidyverse)## Load example datadf <- data.frame(Col1 = LETTERS[1:4], Col2 = LETTERS[23:26], Col3 = c(1:4), Col4 = c(100:103))## Expected outcomedf %>% bind_cols(imap_dfr(df %>% select(Col1, Col4),~ paste(.y, .x, sep = "_")) %>%unite(newcol, sep = "_"))#> Col1 Col2 Col3 Col4 newcol#> 1 A W 1 100 Col1_A_Col4_100#> 2 B X 2 101 Col1_B_Col4_101#> 3 C Y 3 102 Col1_C_Col4_102#> 4 D Z 4 103 Col1_D_Col4_103## With a small number of columnsdf %>%mutate(tmp_Col1 = paste0("Col1", "_", Col1),tmp_Col4 = paste0("Col4", "_", Col4)) %>%unite(newcol, c(tmp_Col1, tmp_Col4), sep = "_")#> Col1 Col2 Col3 Col4 newcol#> 1 A W 1 100 Col1_A_Col4_100#> 2 B X 2 101 Col1_B_Col4_101#> 3 C Y 3 102 Col1_C_Col4_102#> 4 D Z 4 103 Col1_D_Col4_103## With a large number of columnsdf %>%mutate(across(c(Col1, Col4),~paste0(cur_column(), "_", .x))) %>%unite(newcol, c(Col1, Col4), sep = "_") %>%left_join(df)#> Joining with `by = join_by(Col2, Col3)`#> newcol Col2 Col3 Col1 Col4#> 1 Col1_A_Col4_100 W 1 A 100#> 2 Col1_B_Col4_101 X 2 B 101#> 3 Col1_C_Col4_102 Y 3 C 102#> 4 Col1_D_Col4_103 Z 4 D 103
<sup>Created on 2023-06-22 with reprex v2.0.2</sup>
If you have a large number of columns you want to transform, using across() allows you to employ tidyselect functions, such as starts_with(), to select columns of interest without having to specify each column by name.
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论