你如何整理和组合主键略有不同的两个数据框？

Question

流table

experiment	protocol	test	stream_size	metric	value
1	tcp	stream	64	throughput Gbps	10
1	tcp	stream	64	cpu utilization	.5
2	tcp	stream	64	throughput Gbps	40
2	tcp	stream	64	cpu utilization	.9
3	udp	stream	64	throughput Gbps	20
3	udp	stream	64	cpu utilization	.5
4	udp	stream	64	throughput Gbps	60
4	udp	stream	64	cpu utilization	.8

rr table

experiment	protocol	test	request_size	response_size	metric	value
5	tcp	request and response	64	64	transactions per second	10
5	tcp	request and response	64	64	cpu utilization	.6
6	tcp	request and response	64	1024	transactions per second	8
6	tcp	request and response	64	1024	cpu utilization	.5
7	udp	request and response	64	64	transactions per second	30
7	udp	request and response	64	64	cpu utilization	.4
8	udp	request and response	64	1024	transactions per second	29
8	udp	request and response	64	64	cpu utilization	.75

到目前为止，实验的结果列在指标列中，它们的值列在值列中。

我知道我可以删除 stream_size、request_size 和 response_size 等测试特定列，然后绑定行以创建一个数据框。

使用 R 和 tidyverse 工具，您将如何将两个数据框组合成一个长格式，以便组合的数据框没有测试特定列，stream_size，request_size，以及 response_size?

是否有更好或更简洁的方法来为这些实验数据制作模式以方便合并数据框？

Answer 1

您可以将 2 个数据帧绑定在一起，然后仅将以 size 结尾的列转换为长格式。

library(tidyverse)

bind_rows(stream, rr) %>%
  pivot_longer(ends_with("size"), names_to = "test_specific", values_to = "size", values_drop_na = TRUE)

输出

   experiment protocol test                 metric                  value test_specific  size
        <int> <chr>    <chr>                <chr>                   <dbl> <chr>         <int>
 1          1 tcp      stream               throughput Gbps          10   stream_size      64
 2          1 tcp      stream               cpu utilization           0.5 stream_size      64
 3          2 tcp      stream               throughput Gbps          40   stream_size      64
 4          2 tcp      stream               cpu utilization           0.9 stream_size      64
 5          3 udp      stream               throughput Gbps          20   stream_size      64
 6          3 udp      stream               cpu utilization           0.5 stream_size      64
 7          4 udp      stream               throughput Gbps          60   stream_size      64
 8          4 udp      stream               cpu utilization           0.8 stream_size      64
 9          5 tcp      request and response transactions per second  10   request_size     64
10          5 tcp      request and response transactions per second  10   response_size    64
# … with 14 more rows

你如何整理和组合主键略有不同的两个数据框？

how do you tidy and combine two data frames with slightly different primary keys?

r

tidyr