将按行添加的值应用于单行变量,同时保留其他变量和行。

huangapple go评论58阅读模式
英文:

Adding values by row and have them apply to a single row variable while preserving other variables and rows

问题

这是你要的代码部分的翻译:

我有一个看起来像这样的数据框,

df <- data.frame(num1 = c('a','b','c','d')
,num2 = c(1,2,3,4)
,num3 = c(5,6,7,8)
,num4 = c('x','y','b','d'))


我想要的输出是从

num1 num2 num3 num4
a 1 5 x
b 2 6 y
c 3 7 b
d 4 8 d

num1 num2 num3 num4
a 10 26 x
b 2 6 y
c 3 7 b
d 4 8 d


这里是一个可以实现这一结果的示例

df <- data.frame(num1 = c('a', 'b', 'c', 'd'),
num2 = c(1, 2, 3, 4),
num3 = c(5, 6, 7, 8),
num4 = c('x', 'y', 'b', 'd'))

sum_summarised <- df %>%
filter(grepl('a|b|c', num1)) %>%
summarise(num2 = sum(num2), num3 = sum(num3))

df <- df %>%
mutate(num2 = if_else(num1 == 'a', sum_summarised$num2, num2))

df <- df %>%
mutate(num3 = if_else(num1 == 'a', sum_summarised$num3, num3))


基本上是对num2/num3列求和,并将求和结果应用于a行,同时保留变量b、c和d以及num4列的原始行值。

最好使用dplyr - 我尝试过各种`group_by`、`slice`和`filter`的变种组合,但都无济于事。对于我面临的这个独特问题,任何帮助都将不胜感激。谢谢!


<details>
<summary>英文:</summary>

I have a dataframe that looks like,

df <- data.frame(num1 = c('a','b','c','d')
,num2 = c(1,2,3,4)
,num3 = c(5,6,7,8)
,num4 = c('x','y','b','d'))


And would like the out put to go from

num1 num2 num3 num4
a 1 5 x
b 2 6 y
c 3 7 b
d 4 8 d


To

num1 num2 num3 num4
a 10 26 x
b 2 6 y
c 3 7 b
d 4 8 d


Here is a sample that achieves the result in my own solution

df <- data.frame(num1 = c('a', 'b', 'c', 'd'),
num2 = c(1, 2, 3, 4),
num3 = c(5, 6, 7, 8),
num4 = c('x', 'y', 'b', 'd'))

sum_summarised <- df %>%
filter(grepl('a|b|c',num1)) %>%
summarise(num2 = sum(num2), num3 = sum(num3))

df <- df %>%
mutate(num2 = if_else(num1 == 'a',sum_summarised$num2,num2))

df <- df %>%
mutate(num3 = if_else(num1 == 'a',sum_summarised$num3,num3))


Essentially summing num2/num3 columns and applying the sum to row a variable while preserving original row values for variables b,c, and d and num 4 column values.

Preference would be to use dplyr - I have tried variants of `group_by` and `slice` and `filter` combinations to no avail. Any help would be greatly appreciated in this unique problem I have faced. Thank you!

</details>


# 答案1
**得分**: 2

由于您偏好使用 `dplyr`,我们可以使用 `across` 和 `if_else`:

```R
library(dplyr)

df |&gt;
  mutate(across(num2:num3, ~ if_else(num1 == "a", sum(.), .)))

输出:

  num1 num2 num3 num4
1    a   10   26    x
2    b    2    6    y
3    c    3    7    b
4    d    4    8    d

更新 如果我只想对 a、b 和 c 求和,而不是全部,而且只应用于 a 呢?

library(dplyr)

df |&gt;
  mutate(across(num2:num3, ~ if_else(num1 == "a", sum(.[num1 %in% c("a", "b", "c")]), .)))

输出:

  num1 num2 num3 num4
1    a    6   18    x
2    b    2    6    y
3    c    3    7    b
4    d    4    8    d
英文:

Since you have a preference for dplyr, we could use across and if_else:

library(dplyr)

df |&gt;
  mutate(across(num2:num3, ~ if_else(num1 == &quot;a&quot;, sum(.), .)))

Output:

  num1 num2 num3 num4
1    a   10   26    x
2    b    2    6    y
3    c    3    7    b
4    d    4    8    d

Update if i wanted to to just sum a, b, and c? not all? and apply to a?:

library(dplyr)

df |&gt;
  mutate(across(num2:num3, ~ if_else(num1 == &quot;a&quot;, sum(.[num1 %in% c(&quot;a&quot;, &quot;b&quot;, &quot;c&quot;)]), .)))

Output:

  num1 num2 num3 num4
1    a    6   18    x
2    b    2    6    y
3    c    3    7    b
4    d    4    8    d

答案2

得分: 2

另一种方法是使用 rows_update()。这种方法稍微冗长一些,但如果我们想要构建更复杂的操作,我会说它是值得的。下面我们调用 rows_update(),在其中我们使用 summarise() 首先定义我们想要按 id 列进行连接的列,然后使用 across() 更新要更新的列,其他列将保持不变。

library(dplyr)

df %>% 
  rows_update(
    df %>%
      summarise(num1 = "a",
                across(num2:num3, sum)),
    by = "num1")

我们还可以在我们的 tibble 内执行更复杂的操作,例如,如果我们不想对 c 列进行求和,我们可以使用 filter()

df %>% 
  rows_update(
    df %>%
      filter(num1 != "c") %>% 
      summarise(num1 = "a",
                across(num2:num3, sum)),
    by = "num1")

OP 的数据:

df <- data.frame(num1 = c('a','b','c','d'),
                 num2 = c(1,2,3,4),
                 num3 = c(5,6,7,8),
                 num4 = c('x','y','b','d'))

2023-03-03 由 reprex package 创建

英文:

Another approach is to use rows_update(). This is a bit more verbose, but I'd say it pays of if we want to construct more complex operations. Below we call rows_update() and inside we use summarise() first defining the id column we want to join by and then the columns we want to update with `across(), everything else will be untouched.

library(dplyr)

df %&gt;% 
  rows_update(
    df %&gt;%
      summarise(num1 = &quot;a&quot;,
                across(num2:num3, sum)),
    by = &quot;num1&quot;)

#&gt;   num1 num2 num3 num4
#&gt; 1    a   10   26    x
#&gt; 2    b    2    6    y
#&gt; 3    c    3    7    b
#&gt; 4    d    4    8    d

We can also perform more complex operations inside our tibble for example if we don't want to sum-up c we can filter():

df %&gt;% 
  rows_update(
    df %&gt;%
      filter(num1 != &quot;c&quot;) %&gt;% 
      summarise(num1 = &quot;a&quot;,
                across(num2:num3, sum)),
    by = &quot;num1&quot;)

#&gt;   num1 num2 num3 num4
#&gt; 1    a    7   19    x
#&gt; 2    b    2    6    y
#&gt; 3    c    3    7    b
#&gt; 4    d    4    8    d

Data from OP

df &lt;- data.frame(num1 = c(&#39;a&#39;,&#39;b&#39;,&#39;c&#39;,&#39;d&#39;)
                 ,num2 = c(1,2,3,4)
                 ,num3 = c(5,6,7,8)
                 ,num4 = c(&#39;x&#39;,&#39;y&#39;,&#39;b&#39;,&#39;d&#39;))

<sup>Created on 2023-03-03 by the reprex package (v2.0.1)</sup>

答案3

得分: 0

以下是翻译好的代码部分:

# 可能的一种方法是使用'dplyr'包中的`transmute()`函数:
library(tidyverse)
df %>%
  transmute(num1, 
            num2 = c(sum(num2), num2[-1]), 
            num3 = c(sum(num3), num3[-1]), 
            num4)

请注意,这是代码的翻译部分,没有其他内容。

英文:

One way could be using transmute() from 'dplyr' package:

library(tidyverse)
df %&gt;%
  transmute(num1, 
            num2 = c(sum(num2), num2[-1]), 
            num3 = c(sum(num3), num3[-1]), 
            num4)

  num1 num2 num3 num4
1    a   10   26    x
2    b    2    6    y
3    c    3    7    b
4    d    4    8    d

huangapple
  • 本文由 发表于 2023年3月3日 20:56:38
  • 转载请务必保留本文链接:https://go.coder-hub.com/75627371.html
匿名

发表评论

匿名网友

:?: :razz: :sad: :evil: :!: :smile: :oops: :grin: :eek: :shock: :???: :cool: :lol: :mad: :twisted: :roll: :wink: :idea: :arrow: :neutral: :cry: :mrgreen:

确定