删除重复行并计算 R 数据框中的平均值

Question

---

我提供了一个大型数据框的小示例，我正在处理 >1,000 列和 >200 行。我想使每列的每个重复行成为一行而不是两行，同时取与第 1 列相关的每个值的平均值(平均值的列是 Col 6-8 等等) .

---

Col1    Col2    Col3    Col4    Col5    Col6        Col7        Col8        Col9        Col10
A-001   7       40      1       J       3.985645    4.231623    2.36987     9.36545852  8.369663
A-001   7       40      1       J       1.458798    7.652123    1.236985    4.236987    1.22556633
B-002   8       50      0       K       5.00212     8.369562    7.4569852   5.36692     4.6632121
B-002   8       50      0       K       9.02336     1.2120145   3.0014588   8.214569    5.2223698
C-003   10      60      1       L       1.451203    5.321455    8.25963     2.03369878  4.3336988
C-003   10      60      1       L       1.65653     2.369898    8.2136999   7.21458777  5.3366
D-004   3       70      0       M       5.323211    1.147852    7.20014     5.36989     2.36555
D-004   3       70      0       M       4.36969     5.231478    4.23698     3.645478    9.214563
E-005   4       80      1       N       8.123256    9.2356478   5.3696      4.698889    7.366695
E-005   4       80      1       N       7.9632145   0.004555    1.24789     7.3696969   1.23655

预期输出:

col 1   col 2   col 3   col 4   col 5     col 6
A-001       7      40       1       J   2.49282
B-002       8      50       0       K   7.01274
C-003      10      60       1       L   1.55387
D-004       3      70       0       M   4.84645
E-005       4      80       1       N   8.04324

我提前为没有以正确的格式制作它而道歉预先感谢您提供的任何帮助

dd <- structure(list(col1 = structure(c(1L, 1L, 2L, 2L, 3L, 3L, 4L, 
4L, 5L, 5L), .Label = c("A-001", "B-002", "C-003", "D-004", "E-005"
), class = "factor"), col2 = structure(c(1L, 1L, 2L, 2L, 3L, 
3L, 4L, 4L, 5L, 5L), .Label = c("7", "8", "10", "3", "4"), class = "factor"), 
    col3 = c(40L, 40L, 50L, 50L, 60L, 60L, 70L, 70L, 80L, 80L
    ), col4 = c(1, 1, 0, 0, 1, 1, 0, 0, 1, 1), col5 = c(JL, JL, KL, KL, LL, LL, ML, ML, NL, 
    NL), col6 = c(3.985645, 1.458798, 5.00212, 9.02336, 1.451203, 1.65653, 5.323211,  
    4.36969, 8.123256, 7.9632145), col7 = c(4.231623, 7.652123, 8.369562, 1.2120145,  
    5.321455, 2.369898, 1.147852, 5.231478, 9.2356478, 0.004555), col8 = c(2.36987,  
    1.236985, 7.4569852, 3.0014588, 8.25963, 8.2136999, 7.20014, 4.23698, 5.3696,  
    1.24789), col9 = c(9.36545852, 4.236987, 5.36692, 8.214569, 2.03369878, 7.21458777,  
    5.36989, 3.645478, 4.698889, 7.3696969), col10 = c(8.369663, 1.22556633, 4.6632121,  
    5.2223698, 4.3336988, 5.3366, 2.36555, 9.214563, 7.366695, 1.23655)), .Names =  
    c("col1", "col2", "col3", "col4", "col5", "col6", "col7", "col8", "col9", "col10"),  
    class = "data.frame", row.names = c(NA, -10L))

Answer 1

这使用 dplyr ，它对于大数据来说速度很快。第一部分提取未聚合的唯一列，第二部分进行聚合(通过 col1 组)。然后将这两部分绑定(bind)在一起 (bind_cols)。我重命名了您的列以删除空格，如果您保留空格，则需要在名称中使用反引号(即 `col 1`)

library(dplyr)
cols <- 6:10  # columns to average
bind_cols(
    dat %>% distinct(col1) %>% .[,-cols],  # pull out columns we aren't aggregating
    dat[,c(1, cols)] %>% group_by(col1) %>%
        summarise_each(funs(mean)) %>% .[,-1]  # aggregate other columns
)

#    col1 col2 col3 col4 col5      col6       col7      col8      col9      col10
# 1 A-001    7   40    1    J   2.49282    5.94187   1.80343   6.80122    4.79762
# 2 B-002    8   50    0    K   7.01274    4.79079   5.22922   6.79075    4.94279
# 3 C-003   10   60    1    L   1.55387    3.84568   8.23667   4.62414    4.83515
# 4 D-004    3   70    0    M   4.84645    3.18967   5.71856   4.50768    5.79006
# 5 E-005    4   80    1    N   8.04324    4.62010   3.30875   6.03429    4.30162

编辑

这太复杂了，多亏了@StevenBeaupré，简单的版本是(*_each 函数当然可以删除列!)

dat %>% group_by(col1) %>% mutate_each(funs(mean), -(1:5)) %>% distinct