- Java 双重比较
- java - 比较器与 Apache BeanComparator
- Objective-C 完成 block 导致额外的方法调用?
- database - RESTful URI 是否应该公开数据库主键?
我们在夏季收集了超过 130000 份植物物候观测数据,并将数据输入 Excel。每个观察包括 1 到 6 个描述植物物候学不同方面的分类变量。例如,我可能会收集关于一棵白桦树的一项观察结果 - 生长的叶子,或者我可能会收集关于一棵白桦树的两项观察结果 - 叶子生长和开花。
不幸的是,我没有按照数据表上的逻辑顺序收集分类代码,因此在 Excel 中输入了它们,但没有反射(reflect)物候代码的类别(即其他、落叶、开花、水果、叶子),这造成了数据噩梦衰老,叶片脱落)
这是我的数据的样子(在问题底部找到 R 的样本数据):
我的数据应该是这样的:
我创建了一个电子表格,其中包含我所有的物候代码及其相关的物候类别(同样,其他、落叶、开花、果实、叶片衰老、叶片脱落)。
我想使用我导入到 R 中的物候代码电子表格(请参阅底部的代码)将我的数据集重新组织为上面显示的逻辑格式。我可以通过创建每个新字段然后编写大量条件语句(不需要物候代码电子表格!)来做到这一点,但我不知道如何有效地使用我的数据和物候代码来快速有效地重组我的数据。
最后,在我的物候代码电子表格中,我创建了一个排名字段来处理有时我的技术人员在同一类别中记录两个观察结果的事实。在这种情况下,应始终以最高的数字或等级为准。
Sample.Data <- structure(list(Species = c("A", "B", "C", "D", "E","F", "G", "H", "I",
"J", "K", "L", "M", "N", "O", "P", "Q", "R", "S", "T"),
Code.1 = c("C", "C", "C", "C", "C", "C", "C", "C", "C", "C",
"C", "C", "C", "C", "C", "C", "C", "C", "C", "C"),
Code.2 = c("V", "0", "rf", "0", "0", "0", "uf", "uf", "uf", "uf", "0", "0", "0",
"0", "uf", "uf", "0", "0", "0", "0"), Code.3 = c("g3", "gd", "r3", "r3", "r3", "r3",
"V", "V", "V", "V", "g1", "gd", "vd", "g1", "V", "V", "g1", "r3", "r3", "r3"),
Code.4 = c("vd", "vd", "vd", "vd", "vd", "vd", "g3", "g3", "g3", "g3", "vd", "vd", "r2",
"vd", "g1", "vd", "vd", "vd", "vd", "vd"),
Code.5 = c("L2", "L1", "L1", "L2", "L2", "L2", "L2", "L2", "L3", "L2", "L3", "L2", "L2",
"L3", "L1", "L1", "L2", "L1", "L1", "L2"),
Code.6 = c("K", "K", "K", "K", "b1", "b3", "b2", "K", "K", "b4", "K", "K", "K", "b1",
"b3", "Y", "Z", "Y", "K", "b1")), .Names = c("Species", "Code.1", "Code.2",
"Code.3", "Code.4", "Code.5", "Code.6"), row.names = c(NA, -20L), class = "data.frame")
Pheno.Codes <- structure(list(`Pheno Code` = c("Y", "0", "Z", "A", "B1", "B2",
"C", "FA", "As", "Af", "R", "Rs", "Rf", "Ra", "K", "w", "m", "mw",
"wm", "st", "b", "b1", "b2", "b3", "b2", "b4", "uf", "rd", "rf",
"V", "VL", "Vb", "gd", "gb", "g1", "g2", "g3", "ed", "r", "r1",
"r2", "r3", "vd", "vt", "L", "L1", "L2", "L3", "L4", "X"),
`Pheno Category` = c("Other", "Other", "Leaf-out", "Leaf-out",
"Leaf-out", "Leaf-out", "Leaf-out", "Flowering", "Flowering",
"Flowering", "Flowering", "Flowering", "Flowering", "Flowering",
"Flowering", "Flowering", "Flowering", "Flowering", "Flowering",
"Flowering", "Flowering", "Flowering", "Flowering", "Flowering",
"Flowering", "Flowering", "Fruit", "Fruit", "Fruit", "Fruit",
"Fruit", "Fruit", "Leaf senescence", "Leaf senescence",
"Leaf senescence", "Leaf senescence", "Leaf senescence",
"Leaf senescence", "Leaf senescence", "Leaf senescence",
"Leaf senescence", "Leaf senescence", "Leaf senescence",
"Leaf senescence", "Leaf abscission", "Leaf abscission",
"Leaf abscission", "Leaf abscission", "Leaf abscission",
"Other"), Rank = c(0, 0.1, 0.5, 1, 1.1, 1.2, 1.3, 2, 2, 2.1, 2,
2, 2.1, 2.3, 2, 2.1, 2.1, 2.1, 2.1, 2.1, 2.1, 2.1, NA, 2.3, NA,
2.5, 3, 3.1, 3.2, 3.2, 3.2, 3.3, 4, 4, 4.1, 4.2, 4.3, 4.4, 4.4,
4.5, 4.6, 4.7, 4.8, 4.9, 5, 5, 5.1, 5.2, 5.3, 6)), .Names = c("Pheno Code",
"Pheno Category", "Rank"), class = "data.frame", row.names = c(NA, -50L),
class = "data.frame")
Sample.Data2 <- structure(list(Species = c("A", "B", "C", "D", "E","F", "G", "H", "I",
"J", "K", "L", "M", "N", "O", "P", "Q", "R", "S", "T"),
Code.1 = c("C", "C", "B1", "C", "", "C", "C", "C", "C", "C",
"C", "C", "C", "C", "C", "C", "C", "C", "C", "C"),
Code.2 = c("V", "0", "rf", "0", "0", "0", "uf", "uf", "uf", "uf", "0", "", "0",
"0", "uf", "uf", "0", "0", "0", "0"), Code.3 = c("g3", "gd", "r3", "r3", "r3", "r3",
"V", "V", "", "V", "g1", "gd", "vd", "g1", "V", "V", "g1", "r3", "r3", "r3"),
Code.4 = c("", "vd", "vd", "vd", "vd", "vd", "g3", "g3", "g3", "g3", "vd", "vd", "r2",
"qd", "g1", "vd", "vd", "vd", "vd", "vd"),
Code.5 = c("L2", "L1", "L1", "L7", "L2", "L2", "L2", "L2", "L3", "L2", "L3", "L2", "L2",
"L3", "L1", "L1", "L2", "L1", "L1", "L2"),
Code.6 = c("", "", "K", "K", "b1", "b6", "b2", "K", "K", "b4", "K", "K", "K", "b1",
"b3", "Y", "Z", "Y", "K", "b1")), .Names = c("Species", "Code.1", "Code.2",
"Code.3", "Code.4", "Code.5", "Code.6"), row.names = c(NA, -20L), class = "data.frame")
最佳答案
data.table
的可能解决方案:
# load the 'data.table'-package
library(data.table)
# convert both dataframes to data.table's
setDT(Sample.Data)
setDT(Pheno.Codes)
# reshape 'Sample.Data' to long format
sample.long <- melt(Sample.Data, id = 'Species')
# join with 'Pheno.Codes'
# filter/select for each 'Species'/'pheno.cat' combo the row where the rank is equal to the max rank
# reshape the result into wide format again
sample.long[Pheno.Codes, on = c('value' = 'Pheno Code'), `:=` (pheno.cat = `Pheno Category`, rnk = Rank)
][, .SD[rnk == max(rnk)], by = .(Species, pheno.cat)
][, dcast(.SD, Species ~ pheno.cat, value.var = 'value', fill = '')]
给出:
Species Flowering Fruit Leaf abscission Leaf senescence Leaf-out Other
1: A K V L2 vd C
2: B K L1 vd C 0
3: C K rf L1 vd C
4: D K L2 vd C 0
5: E b1 L2 vd C 0
6: F b3 L2 vd C 0
7: G V L2 g3 C
8: H K V L2 g3 C
9: I K V L3 g3 C
10: J b4 V L2 g3 C
11: K K L3 vd C 0
12: L K L2 vd C 0
13: M K L2 vd C 0
14: N b1 L3 vd C 0
15: O b3 V L1 g1 C
16: P V L1 vd C Y
17: Q L2 vd C 0
18: R L1 vd C 0
19: S K L1 vd C 0
20: T b1 L2 vd C 0
更新
为了响应评论中提到的规范,您可以将代码调整为:
setDT(Sample.Data2)
setDT(Pheno.Codes)
sample.long <- melt(Sample.Data2, id = 'Species')[value != '']
sample.long[Pheno.Codes, on = c('value' = 'Pheno Code'), `:=` (pheno.cat = `Pheno Category`, rnk = Rank)
][is.na(pheno.cat), `:=` (pheno.cat = 'ERROR', rnk = 0)
][, .SD[rnk == max(rnk)], by = .(Species, pheno.cat)
][, dcast(.SD, Species ~ pheno.cat, value.var = 'value', fill = '')]
关于r - 如何按类别将列中的数据重组为新列,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/48113251/
我正在从 Stata 迁移到 R(plm 包),以便进行面板模型计量经济学。在 Stata 中,面板模型(例如随机效应)通常报告组内、组间和整体 R 平方。 I have found plm 随机效应
关闭。这个问题不符合Stack Overflow guidelines .它目前不接受答案。 想改进这个问题?将问题更新为 on-topic对于堆栈溢出。 6年前关闭。 Improve this qu
我想要求用户输入整数值列表。用户可以输入单个值或一组多个值,如 1 2 3(spcae 或逗号分隔)然后使用输入的数据进行进一步计算。 我正在使用下面的代码 EXP <- as.integer(rea
当 R 使用分类变量执行回归时,它实际上是虚拟编码。也就是说,省略了一个级别作为基础或引用,并且回归公式包括所有其他级别的虚拟变量。但是,R 选择了哪一个作为引用,以及我如何影响这个选择? 具有四个级
这个问题基本上是我之前问过的问题的延伸:How to only print (adjusted) R-squared of regression model? 我想建立一个线性回归模型来预测具有 15
我在一台安装了多个软件包的 Linux 计算机上安装了 R。现在我正在另一台 Linux 计算机上设置 R。从他们的存储库安装 R 很容易,但我将不得不使用 安装许多包 install.package
我正在阅读 Hadley 的高级 R 编程,当它讨论字符的内存大小时,它说: R has a global string pool. This means that each unique strin
我们可以将 Shiny 代码写在两个单独的文件中,"ui.R"和 "server.R" , 或者我们可以将两个模块写入一个文件 "app.R"并调用函数shinyApp() 这两种方法中的任何一种在性
我正在使用 R 通过 RGP 包进行遗传编程。环境创造了解决问题的功能。我想将这些函数保存在它们自己的 .R 源文件中。我这辈子都想不通怎么办。我尝试过的一种方法是: bf_str = print(b
假设我创建了一个函数“function.r”,在编辑该函数后我必须通过 source('function.r') 重新加载到我的全局环境中。无论如何,每次我进行编辑时,我是否可以避免将其重新加载到我的
例如,test.R 是一个单行文件: $ cat test.R # print('Hello, world!') 我们可以通过Rscript test.R 或R CMD BATCH test.R 来
我知道我可以使用 Rmd 来构建包插图,但想知道是否可以更具体地使用 R Notebooks 来制作包插图。如果是这样,我需要将 R Notebooks 编写为包小插图有什么不同吗?我正在使用最新版本
我正在考虑使用 R 包的共享库进行 R 的站点安装。 多台计算机将访问该库,以便每个人共享相同的设置。 问题是我注意到有时您无法更新包,因为另一个 R 实例正在锁定库。我不能要求每个人都关闭它的 R
我知道如何从命令行启动 R 并执行表达式(例如, R -e 'print("hello")' )或从文件中获取输入(例如, R -f filename.r )。但是,在这两种情况下,R 都会运行文件中
我正在尝试使我当前的项目可重现,因此我正在创建一个主文档(最终是一个 .rmd 文件),用于调用和执行其他几个文档。这样我自己和其他调查员只需要打开和运行一个文件。 当前设置分为三层:主文件、2 个读
关闭。这个问题不符合Stack Overflow guidelines .它目前不接受答案。 想改进这个问题?将问题更新为 on-topic对于堆栈溢出。 5年前关闭。 Improve this qu
我的 R 包中有以下描述文件 Package: blah Title: What the Package Does (one line, title case) Version: 0.0.0.9000
有没有办法更有效地编写以下语句?accel 是一个数据框。 accel[[2]]<- accel[[2]]-weighted.mean(accel[[2]]) accel[[3]]<- accel[[
例如,在尝试安装 R 包时 curl作为 usethis 的依赖项: * installing *source* package ‘curl’ ... ** package ‘curl’ succes
我想将一些软件作为一个包共享,但我的一些脚本似乎并不能很自然地作为函数运行。例如,考虑以下代码块,其中“raw.df”是一个包含离散和连续类型变量的数据框。函数“count.unique”和“squa
我是一名优秀的程序员,十分优秀!