如何在R中添加数据框作为具有匹配名称的顶点属性?

问题描述 投票:0回答:1

在R中,我有如下图“gD”

IGRAPH 40b044a UN-- 17 38 --     
+ attr: name (v/c)    
+ edges from 40b044a (vertex names):    
 [1] Jane     --Jay       Jane     --Brian     Jane     --David     Jane     --Sarah        
 [5] Jane     --Tom       Jay      --Christian Jay      --David     Jay      --Sarah        
 [9] Jay      --Dave      Jay      --Josep     Jay      --Ray       Brian    --David    
[13] Brian    --Sarah     Brian    --Christin  Brian    --Tom       Christian--Sarah    
[17] Christian--Jim       Christian--Dave      Christian--Josep     Michael  --David    
[21] Michael  --Christin  Michael  --Tim       David    --Tim       David    --Tom      
[25] David    --Dave      David    --Zemma     David    --Ray       Jim      --Josep    
[29] Christin --Tom       Christin --Zemma     Tim      --Dickson   Tim      --Zemma    
+ ... omitted several edges    

这是org.unit表,我需要添加以下属性,即grade和org。 org.unit表

    name grade             org
1       Jane    11              HR
2        Tom    11         Finance
3      David     9       Marketing
4        Jay     9       Marketing
5      Brian     8             GTO
6  Christian     7             GTO
7        Tim     5 Commercial Bank

我试图使用set.vertex.attribute(),但我无法弄清楚如何让它通过数据框并只添加属性到现有节点。例如,Jane的等级在org.unit表中是11。我需要先从图dD中检查Jane,然后从org.unit表中分配相应的Jane等级。

当我在下面尝试时,我收到错误消息“在换行期间出错:3个参数传递给'$',需要2”

gD <- gD %>%set_vertex_attr( .,name = 'grade', index = V(gD), value = sapply(V(gD)$name, function(x){org.unit %>% filter( org.unit$name == x) %>% org.unit$grade }))

我花了两天尝试不同的方式,但没有一个工作。请帮忙。

r networking igraph
1个回答
0
投票

您可以设置图表gDfrom表格tab的属性,如下所示:

V(gD)$attribute <- sapply(V(gD)$name, function(x) tab$attribute[tab$virtex.name == x])

小工作代码示例:

这应该模仿您的数据结构:

library(igraph)

# Simple example network similar to your data?
org.unit <- data.frame(name =c("Jane",     "Tom",     "David",       "Jay", "Brian", "Christian",  "Tim"),
                       grade=c(    11,        11,            9,          9,       8,           7,      5),
                       org  =c(  "HR", "Finance", "Marketing", "Marketing",   "GTO",       "GTO", "Bank"))

relations <- data.frame(from=c("Jane",  "Jane",  "Jane",       "Jay",   "Jay", "David"),
                        to = c( "Jay", "Brian", "David", "Christian", "David", "Christian"))

# Make a graph from the relations
gD <- graph.data.frame(relations, directed=TRUE)

# Set virtex atributes
V(gD)$grade <- sapply(V(gD)$name, function(x) org.unit$grade[org.unit$name == x])
V(gD)$org.unit <- sapply(V(gD)$name, function(x) as.character(org.unit$org[org.unit$name == x]))
plot(gD)

# Look at it:
V(gD)$grade

说明:

使用sapply()是愚蠢的,因为它多次将你的org.unit表子集,但它很聪明,因为它保证了顶点的正确顺序。如果你要合并你的表,你的属性的顺序将加扰:

# Using merge() will scramble your vertex order:
attribute_values <- merge(org.unit, data.frame(name=V(gD)$name), by="name")
(attribute_values$name == V(gD)$name)

您当然可以对它们进行排序以强制它们在图表中显示的顺序:

# If you make an attributes table to set from, it would have to be re-ordered
attribute_values <- attribute_values[ match(attribute_values$name, data.frame(name=V(gD)$name)$name), ]
(attribute_values)
# This is the order of your vertices in the graph
(V(gD)$name)

现在你有一个很好的表,按正确的顺序使用你喜欢的方法设置顶点属性。你使用set_vertex_attr()。任何这些都可行:

V(gD)$grade <- attribute_values$grade
gD <- set_vertex_attr(gD, 'grade', V(gD), attribute_values$grade)

请注意:

所有这些代码都假设name是data.frame org.unit的唯一标识符。如果没有,“简”将有多个等级和组织。确认这返回false:

# Are there duplicates in org.unit?
!length(unique(org.unit$name)) == length(org.unit$name)

如果您有多个名为的节点,例如“Jane”,则代码应处理该节点,但会为org.unit中指定的多个Janes分配相同的等级和组织。

祝一切顺利

© www.soinside.com 2019 - 2024. All rights reserved.