如何将列值的下半部分移动到新创建的列中？
How do I move the bottom half of a column's values into a newly created column?
我有一列在前 50% 的行中包含三个不同测量值的平均值，在后 50% 的行中包含相关的标准误差。在前一列中是用于每个的名称（meanNativeSR、meanExoticSR、meanTotalSR、seN、seE、seT）。我想创建 2 个新列，在第一列中包含 se_ 名称，在第二列中包含它们的值，然后删除底部 50% 的行。这是我的数据集：
df <- structure(list(Invasion = structure(c(1L, 1L, 1L, 2L, 2L, 2L,
3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L,
2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L,
1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L
), .Label = c("Low", "Medium", "High"), class = "factor"), Growth = structure(c(1L,
2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L,
3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L,
1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L,
2L, 3L, 1L, 2L, 3L), .Label = c("cover", "herb", "woody"), class = "factor"),
mean_se = c("meanNativeSR", "meanNativeSR", "meanNativeSR",
"meanNativeSR", "meanNativeSR", "meanNativeSR", "meanNativeSR",
"meanNativeSR", "meanNativeSR", "meanExoticSR", "meanExoticSR",
"meanExoticSR", "meanExoticSR", "meanExoticSR", "meanExoticSR",
"meanExoticSR", "meanExoticSR", "meanExoticSR", "meanTotalSR",
"meanTotalSR", "meanTotalSR", "meanTotalSR", "meanTotalSR",
"meanTotalSR", "meanTotalSR", "meanTotalSR", "meanTotalSR",
"seN", "seN", "seN", "seN", "seN", "seN", "seN", "seN", "seN",
"seE", "seE", "seE", "seE", "seE", "seE", "seE", "seE", "seE",
"seT", "seT", "seT", "seT", "seT", "seT", "seT", "seT", "seT"
), value = c(0.769230769230769, 0.230769230769231, 0.923076923076923,
2.46153846153846, 6.84615384615385, 0.538461538461538, 1.69230769230769,
1.76923076923077, 1.15384615384615, 0.384615384615385, 0,
1.38461538461538, 1.76923076923077, 0, 2.23076923076923,
2.07692307692308, 0.769230769230769, 2.46153846153846, 1.15384615384615,
0.230769230769231, 2.53846153846154, 4.23076923076923, 6.84615384615385,
3.23076923076923, 3.76923076923077, 2.76923076923077, 3.84615384615385,
0.280883362823162, 0.12162606385263, 0.329364937914491, 0.312463015562922,
0.705710715103738, 0.24325212770526, 0.36487819155789, 0.280883362823162,
0.191021338791684, 0.140441681411581, 0, 0.180400606147055,
0.201081886427668, 0, 0.230769230769231, 0.329364937914491,
0.12162606385263, 0.24325212770526, 0.273771237231572, 0.12162606385263,
0.24325212770526, 0.394738572265145, 0.705710715103738, 0.440772139427464,
0.532938710021193, 0.257050482766198, 0.336767321450351)), row.names = c(NA,
-54L), class = c("tbl_df", "tbl", "data.frame"))
我能够弄清楚我想用下面的代码做什么，但肯定有更优雅的方法，因为这种方法需要我创建不必要的中间体。
#create an intermediate data.frame that contains just the means and their values from the first half of original df
df.mean <- head(df, -27)
#rename columns 3 and 4
colnames(df.mean)[3] <- "mean"
colnames(df.mean)[4] <- "mean_value"
#create another intermediate data.frame with standard error values from the bottom half of original df
df.se <- df[28:54,]
#rename columns 3 and 4
colnames(df.se)[3] <- "se"
colnames(df.se)[4] <- "se_value"
#cbind those together to get desired result
df.final <- cbind(df.mean, df.se[,3:4])
#remove intermediates
rm(df.mean); rm(df.se)
是否有更简单的方法来完成此操作，也许使用管道或 tidyverse 中的某些函数，或者甚至使用 base R？
这是 pivot_wider 和 unnest 的方法：
library(tidyverse)
df %>%
mutate(class = str_extract(mean_se,"(N|E|T)"),
fun = str_extract(mean_se,"(mean|se)")) %>%
pivot_wider(id_cols = c("Invasion","Growth"), names_from = "fun",
values_from = c("mean_se","value")) %>%
unnest()
# A tibble: 27 x 6
Invasion Growth mean_se_mean mean_se_se value_mean value_se
<fct> <fct> <chr> <chr> <dbl> <dbl>
1 Low cover meanNativeSR seN 0.769 0.281
2 Low cover meanExoticSR seE 0.385 0.140
3 Low cover meanTotalSR seT 1.15 0.274
4 Low herb meanNativeSR seN 0.231 0.122
5 Low herb meanExoticSR seE 0 0
6 Low herb meanTotalSR seT 0.231 0.122
7 Low woody meanNativeSR seN 0.923 0.329
8 Low woody meanExoticSR seE 1.38 0.180
9 Low woody meanTotalSR seT 2.54 0.243
10 Medium cover meanNativeSR seN 2.46 0.312
# … with 17 more rows
您会收到一些警告，但它应该仍然有效。
使用 tidyverse，我们可以做一个 group_split，更改列名，然后做一个 inner_join
library(dplyr)
library(purrr)
df %>%
group_split(grp = row_number() > 27, .keep = FALSE) %>%
map2(list(c('mean', 'mean_value'), c('se', 'se_value')),
~ {nm1 <- .y
.x %>%
rename_at(3:4, ~ nm1) %>%
mutate(rn = row_number())} ) %>%
reduce(inner_join) %>%
select(-rn)
-输出
# A tibble: 27 x 6
# Invasion Growth mean mean_value se se_value
# <fct> <fct> <chr> <dbl> <chr> <dbl>
# 1 Low cover meanNativeSR 0.769 seN 0.281
# 2 Low herb meanNativeSR 0.231 seN 0.122
# 3 Low woody meanNativeSR 0.923 seN 0.329
# 4 Medium cover meanNativeSR 2.46 seN 0.312
# 5 Medium herb meanNativeSR 6.85 seN 0.706
# 6 Medium woody meanNativeSR 0.538 seN 0.243
# 7 High cover meanNativeSR 1.69 seN 0.365
# 8 High herb meanNativeSR 1.77 seN 0.281
# 9 High woody meanNativeSR 1.15 seN 0.191
#10 Low cover meanExoticSR 0.385 seE 0.140
# … with 17 more rows
我认为除了齐心协力之外，没有更短、更容易实现目标的方法。您的代码中最长的部分是分配新的 colnames，这不能真正缩短。其余的可以放在一行中。但实际上，您必须始终平衡简洁性和可读性。
上面显示的 dplyr 方法非常简洁，但我相信它们可以处理比您处理的更多 complex/general 个案例。
df_final_2 <- cbind(head(df, -27), df[28:54,3:4])
colnames(df_final_2)[3:6] <- c("mean", "mean_value","se", "se_value")
我有一列在前 50% 的行中包含三个不同测量值的平均值，在后 50% 的行中包含相关的标准误差。在前一列中是用于每个的名称（meanNativeSR、meanExoticSR、meanTotalSR、seN、seE、seT）。我想创建 2 个新列，在第一列中包含 se_ 名称，在第二列中包含它们的值，然后删除底部 50% 的行。这是我的数据集：
df <- structure(list(Invasion = structure(c(1L, 1L, 1L, 2L, 2L, 2L,
3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L,
2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L,
1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L
), .Label = c("Low", "Medium", "High"), class = "factor"), Growth = structure(c(1L,
2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L,
3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L,
1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L,
2L, 3L, 1L, 2L, 3L), .Label = c("cover", "herb", "woody"), class = "factor"),
mean_se = c("meanNativeSR", "meanNativeSR", "meanNativeSR",
"meanNativeSR", "meanNativeSR", "meanNativeSR", "meanNativeSR",
"meanNativeSR", "meanNativeSR", "meanExoticSR", "meanExoticSR",
"meanExoticSR", "meanExoticSR", "meanExoticSR", "meanExoticSR",
"meanExoticSR", "meanExoticSR", "meanExoticSR", "meanTotalSR",
"meanTotalSR", "meanTotalSR", "meanTotalSR", "meanTotalSR",
"meanTotalSR", "meanTotalSR", "meanTotalSR", "meanTotalSR",
"seN", "seN", "seN", "seN", "seN", "seN", "seN", "seN", "seN",
"seE", "seE", "seE", "seE", "seE", "seE", "seE", "seE", "seE",
"seT", "seT", "seT", "seT", "seT", "seT", "seT", "seT", "seT"
), value = c(0.769230769230769, 0.230769230769231, 0.923076923076923,
2.46153846153846, 6.84615384615385, 0.538461538461538, 1.69230769230769,
1.76923076923077, 1.15384615384615, 0.384615384615385, 0,
1.38461538461538, 1.76923076923077, 0, 2.23076923076923,
2.07692307692308, 0.769230769230769, 2.46153846153846, 1.15384615384615,
0.230769230769231, 2.53846153846154, 4.23076923076923, 6.84615384615385,
3.23076923076923, 3.76923076923077, 2.76923076923077, 3.84615384615385,
0.280883362823162, 0.12162606385263, 0.329364937914491, 0.312463015562922,
0.705710715103738, 0.24325212770526, 0.36487819155789, 0.280883362823162,
0.191021338791684, 0.140441681411581, 0, 0.180400606147055,
0.201081886427668, 0, 0.230769230769231, 0.329364937914491,
0.12162606385263, 0.24325212770526, 0.273771237231572, 0.12162606385263,
0.24325212770526, 0.394738572265145, 0.705710715103738, 0.440772139427464,
0.532938710021193, 0.257050482766198, 0.336767321450351)), row.names = c(NA,
-54L), class = c("tbl_df", "tbl", "data.frame"))
我能够弄清楚我想用下面的代码做什么，但肯定有更优雅的方法，因为这种方法需要我创建不必要的中间体。
#create an intermediate data.frame that contains just the means and their values from the first half of original df
df.mean <- head(df, -27)
#rename columns 3 and 4
colnames(df.mean)[3] <- "mean"
colnames(df.mean)[4] <- "mean_value"
#create another intermediate data.frame with standard error values from the bottom half of original df
df.se <- df[28:54,]
#rename columns 3 and 4
colnames(df.se)[3] <- "se"
colnames(df.se)[4] <- "se_value"
#cbind those together to get desired result
df.final <- cbind(df.mean, df.se[,3:4])
#remove intermediates
rm(df.mean); rm(df.se)
是否有更简单的方法来完成此操作，也许使用管道或 tidyverse 中的某些函数，或者甚至使用 base R？
这是 pivot_wider 和 unnest 的方法：
library(tidyverse)
df %>%
mutate(class = str_extract(mean_se,"(N|E|T)"),
fun = str_extract(mean_se,"(mean|se)")) %>%
pivot_wider(id_cols = c("Invasion","Growth"), names_from = "fun",
values_from = c("mean_se","value")) %>%
unnest()
# A tibble: 27 x 6
Invasion Growth mean_se_mean mean_se_se value_mean value_se
<fct> <fct> <chr> <chr> <dbl> <dbl>
1 Low cover meanNativeSR seN 0.769 0.281
2 Low cover meanExoticSR seE 0.385 0.140
3 Low cover meanTotalSR seT 1.15 0.274
4 Low herb meanNativeSR seN 0.231 0.122
5 Low herb meanExoticSR seE 0 0
6 Low herb meanTotalSR seT 0.231 0.122
7 Low woody meanNativeSR seN 0.923 0.329
8 Low woody meanExoticSR seE 1.38 0.180
9 Low woody meanTotalSR seT 2.54 0.243
10 Medium cover meanNativeSR seN 2.46 0.312
# … with 17 more rows
您会收到一些警告，但它应该仍然有效。
使用 tidyverse，我们可以做一个 group_split，更改列名，然后做一个 inner_join
library(dplyr)
library(purrr)
df %>%
group_split(grp = row_number() > 27, .keep = FALSE) %>%
map2(list(c('mean', 'mean_value'), c('se', 'se_value')),
~ {nm1 <- .y
.x %>%
rename_at(3:4, ~ nm1) %>%
mutate(rn = row_number())} ) %>%
reduce(inner_join) %>%
select(-rn)
-输出
# A tibble: 27 x 6
# Invasion Growth mean mean_value se se_value
# <fct> <fct> <chr> <dbl> <chr> <dbl>
# 1 Low cover meanNativeSR 0.769 seN 0.281
# 2 Low herb meanNativeSR 0.231 seN 0.122
# 3 Low woody meanNativeSR 0.923 seN 0.329
# 4 Medium cover meanNativeSR 2.46 seN 0.312
# 5 Medium herb meanNativeSR 6.85 seN 0.706
# 6 Medium woody meanNativeSR 0.538 seN 0.243
# 7 High cover meanNativeSR 1.69 seN 0.365
# 8 High herb meanNativeSR 1.77 seN 0.281
# 9 High woody meanNativeSR 1.15 seN 0.191
#10 Low cover meanExoticSR 0.385 seE 0.140
# … with 17 more rows
我认为除了齐心协力之外，没有更短、更容易实现目标的方法。您的代码中最长的部分是分配新的 colnames，这不能真正缩短。其余的可以放在一行中。但实际上，您必须始终平衡简洁性和可读性。
上面显示的 dplyr 方法非常简洁，但我相信它们可以处理比您处理的更多 complex/general 个案例。
df_final_2 <- cbind(head(df, -27), df[28:54,3:4])
colnames(df_final_2)[3:6] <- c("mean", "mean_value","se", "se_value")