在R中，根据几列名称中的特定前缀重新排序矩阵列

Question

我在R中有一个带列名的矩阵。

> colnames(m)
 [1] "caz_RNAi1_R1"     "caz_RNAi2_R1"     "cg1316_RNAi1_R1"  "cg1316_RNAi2_R1"  "cg4612_RNAi1_R1" 
 [6] "cg4612_RNAi2_R1"  "Dp1_RNAi1_R1"     "Dp1_RNAi2_R1"     "fmr1_RNAi1_R1"    "fmr1_RNAi2_R1"   
[11] "GFP_RNAi1_R1"     "GFP_RNAi2_R1"     "GFP_RNAi3_R1"     "GFP_RNAi4_R1"     "GFP_RNAi5_R1"    
[16] "GFP_RNAi6_R1"     "hrb87f_RNAi1_R1"  "hrb87f_RNAi2_R1"  "hrb98de_RNAi1_R1" "hrb98de_RNAi2_R1"

现在，一些列名有前缀GFP。我想重新排列矩阵的列，使得那些在名称中具有此前缀的列成为开始的列，其余的列在它们自己之间按字母顺序排序。

因此，colnames(m)应该像这样排序：

"GFP_a", "GFP_b", "GFP_c",..."GFP_z", "a", "b","c","d", ....

如何完成这个任务？

- hi15

---

你想要一个混合排序，可以对字符和数字进行排序吗？ - David Arenburg

Answer 1

你可以这样做：

 m[order(-(grepl('^GFP', m))+1L)]

其中m来自@Mamoun Benghezal的帖子。在此示例中，它已按字母顺序排序，但如果未排序，则

 set.seed(24)
 m1 <-sample(m)
 m1[order(m1)][order(-(grepl('^GFP',m1[order(m1)]))+1L)]

- akrun

Answer 2

你可以尝试这个

m <-  c("caz_RNAi1_R1", "caz_RNAi2_R1", "cg1316_RNAi1_R1", "cg1316_RNAi2_R1", "cg4612_RNAi1_R1",
        "cg4612_RNAi2_R1", "Dp1_RNAi1_R1", "Dp1_RNAi2_R1", "fmr1_RNAi1_R1", "fmr1_RNAi2_R1",
        "GFP_RNAi1_R1", "GFP_RNAi2_R1", "GFP_RNAi3_R1", "GFP_RNAi4_R1", "GFP_RNAi5_R1",
        "GFP_RNAi6_R1", "hrb87f_RNAi1_R1",  "hrb87f_RNAi2_R1",  "hrb98de_RNAi1_R1", "hrb98de_RNAi2_R1")
sort(m[grep(pattern="^GFP", x = m )]) # beginning with GFP
## [1] "GFP_RNAi1_R1" "GFP_RNAi2_R1" "GFP_RNAi3_R1" "GFP_RNAi4_R1" "GFP_RNAi5_R1" "GFP_RNAi6_R1"
sort(m[-grep(pattern="^GFP", x = m )]) # do not begin by GFP
##     [1] "caz_RNAi1_R1"     "caz_RNAi2_R1"     "cg1316_RNAi1_R1"  "cg1316_RNAi2_R1"  "cg4612_RNAi1_R1"  "cg4612_RNAi2_R1"  "Dp1_RNAi1_R1"     "Dp1_RNAi2_R1"     "fmr1_RNAi1_R1"    "fmr1_RNAi2_R1"   
##    [11] "hrb87f_RNAi1_R1"  "hrb87f_RNAi2_R1"  "hrb98de_RNAi1_R1" "hrb98de_RNAi2_R1"
c(sort(m[grep(pattern="^GFP", x = m )]),  sort(m[-grep(pattern="^GFP", x = m )])) # ordered columns

- Mamoun Benghezal

---

这里没有排序元素。 - David Arenburg

即使在这个例子中没有必要，sort也会对列进行排序。 - Mamoun Benghezal