如何使用过滤器功能来提高样品的同质性？答案

【问题标题】：How can I use the filter function to increase sample homogeniety?如何使用过滤器功能来提高样品的同质性？
【发布时间】：2021-08-02 07:54:57
【问题描述】：

我正在寻找一些数据争论的建议。我的数据集包含两组投资者（签字人=0，签字人=1）。都有对应的国家，但是两组的国家不匹配。

对于我的下一个分析，我需要将我的数据集减少到仅存在于两组中的国家/地区，因此每个组将在所列国家/地区中至少有一个单位（投资者）。

需要明确的是，如果一组在 45 个国家/地区有投资者，而另一组在 50 个国家/地区有投资者，但这些国家中只有 30 个匹配，我只想在新数据框中保留这 30 个匹配的国家。

我的数据如下所示：

investor	year	activity	country	region	strategy	signatory
123 IM	2002	4.45	France	europe	VC	1
123 IM	2003	3.2	France	europe	VC	1
123 IM	2004	7.8	France	europe	VC	1
21Invest	2002	4.45	France	europe	VC	0
21Invest	2003	3.2	France	europe	VC	0
21Invest	2004	7.8	France	europe	VC	0
Aegon	2005	5.4	Netherlands	europe	BY	1
Aegon	2006	4.2	Netherlands	europe	BY	1
Aegon	2007	1.3	Netherlands	europe	BY	1
ING	2005	5.4	Netherlands	europe	BY	0
ING	2006	4.2	Netherlands	europe	BY	0
ING	2007	1.3	Netherlands	europe	BY	0
aberdeen	2002	4.45	UK	europe	VC	1
aberdeen	2003	3.2	UK	europe	VC	1
aberdeen	2004	7.8	UK	europe	VC	1
JPM	2005	5.4	USA	europe	BY	0
JPM	2006	4.2	USA	europe	BY	0
JPM	2007	1.3	USA	europe	BY	0

我正在寻找的输出是：

investor	year	activity	country	region	strategy	signatory
123 IM	2002	4.45	France	europe	VC	1
123 IM	2003	3.2	France	europe	VC	1
123 IM	2004	7.8	France	europe	VC	1
21Invest	2002	4.45	France	europe	VC	0
21Invest	2003	3.2	France	europe	VC	0
21Invest	2004	7.8	France	europe	VC	0
Aegon	2005	5.4	Netherlands	europe	BY	1
Aegon	2006	4.2	Netherlands	europe	BY	1
Aegon	2007	1.3	Netherlands	europe	BY	1
ING	2005	5.4	Netherlands	europe	BY	0
ING	2006	4.2	Netherlands	europe	BY	0
ING	2007	1.3	Netherlands	europe	BY	0

注意：英国和美国的公司被删除，而法国和荷兰的公司被保留。

这是因为两个投资者样本（签署人 = 0 和签署人 = 1）在法国/荷兰国家/地区都有单位，而英国和美国仅在其中一个样本中出现。

df <- data.frame(
investor=c("123 IM", "123 IM", "123 IM", "21Invest", "21Invest", "21Invest", "Aegon", "Aegon", "Aegon", "ING", "ING", "ING", "aberdeen", "aberdeen", "aberdeen", "JPM", "JPM", "JPM"), year=c(2002, 2003, 2004, 2005, 2006, 2007, 2002, 2003, 2004, 2005, 2006, 2007, 2002, 2003, 2004, 2005, 2006, 2007),
activity=c(4.45, 3.2, 7.8, 5.4, 4.2, 1.3, 4.45, 3.2, 7.8, 5.4, 4.2, 1.3, 4.45, 3.2, 7.8, 5.4, 4.2, 1.3),
country=c("France", "France", "France", "France", "France", "France", "Netherlands", "Netherlands", "Netherlands", "Netherlands", "Netherlands", "Netherlands", "UK", "UK", "UK", "USA", "USA", "USA"),
region=c("europe", "europe", "europe", "europe", "europe", "europe", "europe", "europe", "europe", "europe", "europe", "europe", "europe", "europe", "europe", "north america", "north america", "north america"),
strategy =c("VC", "VC", "VC", "BY", "BY", "BY", "VC", "VC", "VC", "BY", "BY", "BY", "VC", "VC", "VC", "BY", "BY", "BY"),
signatory =c(1, 1, 1, 0, 0, 0, 1, 1, 1, 0, 0, 0, 1, 1, 1, 0, 0, 0))

df <- data.frame(
investor=c("123 IM", "123 IM", "123 IM", "21Invest", "21Invest", "21Invest", "Aegon", "Aegon", "Aegon", "ING", "ING", "ING", "aberdeen"), year=c(2002, 2003, 2004, 2005, 2006, 2007, 2002, 2003, 2004, 2005, 2006, 2007),
activity=c(4.45, 3.2, 7.8, 5.4, 4.2, 1.3, 4.45, 3.2, 7.8, 5.4, 4.2, 1.3),
country=c("France", "France", "France", "France", "France", "France", "Netherlands", "Netherlands", "Netherlands", "Netherlands", "Netherlands", "Netherlands"),
region=c("europe", "europe", "europe", "europe", "europe", "europe", "europe", "europe", "europe", "europe", "europe", "europe"),
strategy =c("VC", "VC", "VC", "BY", "BY", "BY", "VC", "VC", "VC", "BY", "BY", "BY"),
signatory =c(1, 1, 1, 0, 0, 0, 1, 1, 1, 0, 0, 0))

任何提示将不胜感激！

罗里

【问题讨论】：

您的示例数据框会建议您最终没有行，因为所有签署方都在法国，所有非签署方都在荷兰。根本没有重叠。你能提供更好的数据集吗？
您好，刚刚按照您的评论进行了修改！
是的，看到了。也修改了我的答案。
如果我在下面的回答回答了您的问题，请点击旁边的 v 标志接受答案。您可以对您之前提出的已回答的其他问题执行相同的操作。如果您发现其中任何一个有帮助，您也可以点赞。
嗨 coffieinjunky，感谢您迄今为止的帮助，它改进了数据，但似乎效果并不理想 - 我仍然有 3 个国家/地区出现在数据集中，只有 1 个投资者。知道这是怎么发生的吗？

标签： r filter data-wrangling

【解决方案1】：

你可以这样做：

library(tidyverse)

signatory_countries <- unique(df[df$signatory==1, 'country'])
non_signatory_countries <- unique(df[df$signatory==0, 'country'])

new_df <- bind_rows(
  df %>% filter(signatory==1, country %in% non_signatory_countries),
  df %>% filter(signatory==0, country %in% signatory_countries)
)
new_df
   investor year activity     country region strategy signatory
1    123 IM 2002     4.45      France europe       VC         1
2    123 IM 2003     3.20      France europe       VC         1
3    123 IM 2004     7.80      France europe       VC         1
4     Aegon 2002     4.45 Netherlands europe       VC         1
5     Aegon 2003     3.20 Netherlands europe       VC         1
6     Aegon 2004     7.80 Netherlands europe       VC         1
7  21Invest 2005     5.40      France europe       BY         0
8  21Invest 2006     4.20      France europe       BY         0
9  21Invest 2007     1.30      France europe       BY         0
10      ING 2005     5.40 Netherlands europe       BY         0
11      ING 2006     4.20 Netherlands europe       BY         0
12      ING 2007     1.30 Netherlands europe       BY         0

【讨论】：