Question

是否可以重新排序数据帧的列，这是dcast（）调用的结果 E.x。

鉴于数据：

> dput(copyOfRes)
structure(list(docName = c("doc2", "doc1", "doc1", "doc1", "doc1", 
"doc1", "doc1", "doc1", "doc1", "doc1", "doc1", "doc2"), day_of_week = c(11, 
11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 2)), .Names = c("docName", 
"week_number"), row.names = c(NA, -12L), class = "data.frame")

所以，当我使用dcast（）时如下：

library(reshape2)
dcast(copyOfRes, docName ~ week_number, length)

结果是：

  docName 2 11
1    doc1 0 10
2    doc2 1  1

我希望数据框的值week_number递减，如下所示：

  docName 11  2
1    doc1 10 0
2    doc2 1  1

我尝试过dcast(copyOfRes, docName ~ sort(week_number, decreasing= TRUE), length)，但它仍然不起作用。有什么建议吗？

Answer 1

您可以在factor()内dcast()设置适当的关卡顺序。

 dcast(copyOfRes, 
   docName ~ factor(week_number,levels=unique(week_number)), length)
      docName 11 2
    1    doc1 10 0
    2    doc2  1 1

Answer 2

您可以reorder使用rev

dcast(copyOfRes, docName ~ reorder(week_number,rev(week_number)), length)
Using week_number as value column: use value.var to override.
  docName 11 2
1    doc1 10 0
2    doc2  1 1

重新排序dcast数据框

2 个答案: