有没有办法在 R 中将数据透视_更长到多个值列?
我正在尝试使用pivot_longer来延长我的数据框,但我不需要它很长,并且希望输出多个“值”列。
示例:
df <- tibble(
ids = c("protein1", "protein2"),
mean.group1 = sample(1:1000, 2),
mean.group2 = sample(1:1000, 2),
se.group1 = sample(1:10, 2),
se.group2 = sample(1:10, 2)
)
df
# A tibble: 2 × 5
ids mean.group1 mean.group2 se.group1 se.group2
<chr> <int> <int> <int> <int>
1 protein1 763 456 6 4
2 protein2 820 624 4 7
我想要的输出是:
df2 <- tibble(
ids = c("protein1", "protein1", "protein2", "protein2"),
mean = c(df$mean.group1[1], df$mean.group2[1], df$mean.group1[2], df$mean.group2[2]),
se = c(df$se.group1[1], df$se.group2[1], df$se.group1[2], df$se.group2[2]),
group = c("group1", "group2", "group1", "group2")
)
df2
# A tibble: 4 × 4
ids mean se group
<chr> <int> <int> <chr>
1 protein1 763 6 group1
2 protein1 456 4 group2
3 protein2 820 4 group1
4 protein2 624 7 group2
到目前为止,我已经尝试了多个后续的 pivot_longer()
,然后是 unique()
,但这搞乱了输出:
df_longer <- df %>%
pivot_longer(cols = starts_with("mean."),
names_to = "group",
names_prefix = "mean.",
values_to = "mean") %>%
unique() %>%
pivot_longer(cols = starts_with("se."),
names_to = "group",
names_prefix = "se.",
values_to = "se",
names_repair = "unique") %>%
unique()
df_longer
# A tibble: 8 × 5
ids group...2 mean group...4 se
<chr> <chr> <int> <chr> <int>
1 protein1 group1 763 group1 6
2 protein1 group1 763 group2 4
3 protein1 group2 456 group1 6
4 protein1 group2 456 group2 4
5 protein2 group1 820 group1 4
6 protein2 group1 820 group2 7
7 protein2 group2 624 group1 4
8 protein2 group2 624 group2 7
我有点理解为什么 - 行被重复太多次,因此没有为每行保留组标识。但是,我很难找到解决方案。我知道有一个 names_pattern
选项,但我不确定它在这种情况下如何应用。
任何帮助将不胜感激!我考虑过转换为全长格式(即为每个“平均值”、“se”等设置一个“测量”列),然后使用 pivot_wider()
转换为我需要的格式,但我也不知道该怎么做。另外,如果需要更多信息,请告诉我。我的实际数据集处理4种不同的测量(相同格式,即measurement.group)和数千种蛋白质,但原理应该是相同的,我希望!
I'm trying to use pivot_longer to enlongate my dataframe, but I don't need it to be fully long, and would like to output multiple "values" columns.
Example:
df <- tibble(
ids = c("protein1", "protein2"),
mean.group1 = sample(1:1000, 2),
mean.group2 = sample(1:1000, 2),
se.group1 = sample(1:10, 2),
se.group2 = sample(1:10, 2)
)
df
# A tibble: 2 × 5
ids mean.group1 mean.group2 se.group1 se.group2
<chr> <int> <int> <int> <int>
1 protein1 763 456 6 4
2 protein2 820 624 4 7
My desired output is:
df2 <- tibble(
ids = c("protein1", "protein1", "protein2", "protein2"),
mean = c(df$mean.group1[1], df$mean.group2[1], df$mean.group1[2], df$mean.group2[2]),
se = c(df$se.group1[1], df$se.group2[1], df$se.group1[2], df$se.group2[2]),
group = c("group1", "group2", "group1", "group2")
)
df2
# A tibble: 4 × 4
ids mean se group
<chr> <int> <int> <chr>
1 protein1 763 6 group1
2 protein1 456 4 group2
3 protein2 820 4 group1
4 protein2 624 7 group2
So far, I have tried multiple subsequent pivot_longer()
followed by unique()
, but this is messing up the output:
df_longer <- df %>%
pivot_longer(cols = starts_with("mean."),
names_to = "group",
names_prefix = "mean.",
values_to = "mean") %>%
unique() %>%
pivot_longer(cols = starts_with("se."),
names_to = "group",
names_prefix = "se.",
values_to = "se",
names_repair = "unique") %>%
unique()
df_longer
# A tibble: 8 × 5
ids group...2 mean group...4 se
<chr> <chr> <int> <chr> <int>
1 protein1 group1 763 group1 6
2 protein1 group1 763 group2 4
3 protein1 group2 456 group1 6
4 protein1 group2 456 group2 4
5 protein2 group1 820 group1 4
6 protein2 group1 820 group2 7
7 protein2 group2 624 group1 4
8 protein2 group2 624 group2 7
I sort of understand why - the rows are being duplicated too many times, and thus the group identity is not being kept for each row. However, I'm having trouble coming up with a solution. I'm aware that there's a names_pattern
option but I'm not sure how it would apply in this case.
Any help would be much appreciated! I've considered pivoting to fully long format (i.e. having a "measurement" column for each 'mean', 'se', etc) and then using pivot_wider()
to pivot to the format I need, but I also haven't been able to figure out how to do that. As well, let me know if any more information is needed. My actual dataset deals with 4 different measurements (same format, i.e. measurement.group) and thousands of proteins, but the principle should be the same, I hope!
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(1)
如果我们将
names_to
指定为值向量,即.value
- 返回列的值并使用后缀对列进行“分组”,则不需要多次调用列名称。在这里,我们使用names_sep
作为.
在.
处进行分割-output
注意:值与
sample
不同用于创建没有指定set.seed
的输入数据We don't need multiple calls if we specify the
names_to
as a vector of values i.e..value
- returns the value of the columns and 'group' the column with the suffix of column names. Here, we usenames_sep
as.
to split at the.
-output
NOTE: values are different as
sample
was used in creation of input data without aset.seed
specified