根据另一个数据集获取数据集的子集 [英] take subset of dataset based on another dataset
问题描述
假设我有一个数据集，它是dat1
ID块图SPID TotHeight
1 1 1 4 44.5
2 1 1 4 51
3 1 1 4 28.7
4 1 1 4 24.5
5 1 1 4 27.3
6 1 1 4 20
17 1 10 1 44.5
19 1 10 1 51
1 1 11 21 28.7
2 1 11 21 24.5
3 1 11 21 27.3
4 1 11 21 20
5 1 11 21 12.88666667
6 1 11 21 7.235238095
7 1 11 21 1.583809524
然后我有另一个大数据集，它是dat2：
ID块图SPID种类TotHeight
1 1 1 4 BENI 72
2 1 1 4 BENI 55
3 1 1 4 BENI 51
4 1 1 4 BENI 47
5 1 1 4 BENI 49
6 1 1 4 BENI 34
7 1 1 4 BENI。
8 1 1 4 BENI 51
9 1 1 4 BENI 66
10 1 1 4 BENI 40
11 1 1 4 BENI 24
12 1 1 4 BENI 62
13 1 1 4 BENI 34
14 1 1 4 BENI 49
15 1 1 4 BENI 57
16 1 1 4 BENI 22
17 1 1 4 BENI 76
18 1 1 4 BENI 56
19 1 1 4 BENI 55
20 1 1 4 BENI 29
21 1 1 4 BENI 24
22 1 1 4 BENI 18
23 1 1 4 BENI 65
24 1 1 4 BENI 55
25 1 1 4 BENI 63
26 1 1 4 BENI 57
27 1 1 4 BENI 57
28 1 1 4 BENI 57
29 1 1 4 BENI 45
30 1 1 4 BENI 83
31 1 1 4 BENI 37
32 1 1 4 BENI 56
33 1 1 4 BENI 65
34 1 1 4 BENI 75
35 1 1 4 BENI 51
36 1 1 4 BENI。
1 1 2 16 PRSE 141
2 1 2 16 PRSE 192
3 1 2 16 PRSE。
4 1 2 16 PRSE 197
5 1 2 16 PRSE 172
6 1 2 16 PRSE 143
7 1 2 16 PRSE 141
8 1 2 16 PRSE 155
9 1 2 16 PRSE 167
10 1 2 16 PRSE 155
11 1 2 16 PRSE 175
12 1 2 16 PRSE 190
13 1 2 16 PRSE 148
14 1 2 16 PRSE 180
15 1 2 16 PRSE。
我的问题是如何从dat2中获取数据的一个子集，其中ID，块和图匹配那些在dat1？如何从dat2获取数据的一个子集，其中ID，块和图与dat1不一致？
这听起来像你想要在列ID，块和图上的合并。
假设你的数据被读取在 dat1 和 dat2 中，这应该是你想要的：
>合并（dat1，dat2，by = c（ID，block，plot））
ID块图SPID.x TotHeight.x SPID.y种类TotHeight.y
1 1 1 1 4 44.5 4 BENI 72
2 2 1 1 4 51.0 4 BENI 55
3 3 1 1 4 28.7 4 BENI 51
4 4 1 1 4 24.5 4 BENI 47
5 5 1 1 4 27.3 4 BENI 49
6 6 1 1 4 20.0 4 BENI 34
这实际上是对三个关注的列中的SQL术语执行内部连接。阅读帮助页面合并左，右和外连接的可能性，如果这也是感兴趣的。
完全可重复的要点这里
要获取未从 dat2合并的行，这个黑客的作品。这可能是一个更有效的方法，但是在这里。首先，添加参数 all.y = TRUE 。这指定一个正确的连接，它将从未合并的 dat2 返回行。然后，我们可以将子集，知道没有合并的行将返回 NA 的：
<$ p $子集（dat1，dat2，by = c（ID，block，plot），all.y = TRUE），is.na（SPID.x）== TRUE）
suppose I have one dataset, which is dat1
ID block plot SPID TotHeight
1 1 1 4 44.5
2 1 1 4 51
3 1 1 4 28.7
4 1 1 4 24.5
5 1 1 4 27.3
6 1 1 4 20
17 1 10 1 44.5
19 1 10 1 51
1 1 11 21 28.7
2 1 11 21 24.5
3 1 11 21 27.3
4 1 11 21 20
5 1 11 21 12.88666667
6 1 11 21 7.235238095
7 1 11 21 1.583809524
Then I have another big dataset, which is dat2:
ID block plot SPID Species TotHeight
1 1 1 4 BENI 72
2 1 1 4 BENI 55
3 1 1 4 BENI 51
4 1 1 4 BENI 47
5 1 1 4 BENI 49
6 1 1 4 BENI 34
7 1 1 4 BENI .
8 1 1 4 BENI 51
9 1 1 4 BENI 66
10 1 1 4 BENI 40
11 1 1 4 BENI 24
12 1 1 4 BENI 62
13 1 1 4 BENI 34
14 1 1 4 BENI 49
15 1 1 4 BENI 57
16 1 1 4 BENI 22
17 1 1 4 BENI 76
18 1 1 4 BENI 56
19 1 1 4 BENI 55
20 1 1 4 BENI 29
21 1 1 4 BENI 24
22 1 1 4 BENI 18
23 1 1 4 BENI 65
24 1 1 4 BENI 55
25 1 1 4 BENI 63
26 1 1 4 BENI 57
27 1 1 4 BENI 57
28 1 1 4 BENI 57
29 1 1 4 BENI 45
30 1 1 4 BENI 83
31 1 1 4 BENI 37
32 1 1 4 BENI 56
33 1 1 4 BENI 65
34 1 1 4 BENI 75
35 1 1 4 BENI 51
36 1 1 4 BENI .
1 1 2 16 PRSE 141
2 1 2 16 PRSE 192
3 1 2 16 PRSE .
4 1 2 16 PRSE 197
5 1 2 16 PRSE 172
6 1 2 16 PRSE 143
7 1 2 16 PRSE 141
8 1 2 16 PRSE 155
9 1 2 16 PRSE 167
10 1 2 16 PRSE 155
11 1 2 16 PRSE 175
12 1 2 16 PRSE 190
13 1 2 16 PRSE 148
14 1 2 16 PRSE 180
15 1 2 16 PRSE .
My question is how can I take a subset of data from dat2 in which ID, block and plot match those in dat1? And how can I get a subset of data from dat2 in which ID, block and plot do not match those in dat1?
It sounds like you want a merge on the columns ID, block, and plot.
Assuming your data are read in named dat1 and dat2, this should be what you want:
> merge(dat1, dat2, by = c("ID", "block", "plot"))
ID block plot SPID.x TotHeight.x SPID.y Species TotHeight.y
1 1 1 1 4 44.5 4 BENI 72
2 2 1 1 4 51.0 4 BENI 55
3 3 1 1 4 28.7 4 BENI 51
4 4 1 1 4 24.5 4 BENI 47
5 5 1 1 4 27.3 4 BENI 49
6 6 1 1 4 20.0 4 BENI 34
This essentially performs an inner join in SQL terms on the three columns of interest. Read the help page for merge for left, right, and outer join possibilities if that is also of interest.
Fully reproducible gist here
To get the rows that didn't merge from dat2, this hack works. There is probably a more efficient way to do this, but here it is. First, add the parameter all.y = TRUE. This specifies a right join which will return rows from dat2 which did not merge. Then we can subset on that knowing that the rows that didn't merge will return NA's:
subset(merge(dat1, dat2, by = c("ID", "block", "plot"), all.y = TRUE), is.na(SPID.x) == TRUE)
这篇关于根据另一个数据集获取数据集的子集的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
