[英]finding the overleap values in R
d1<-data.frame(ID=c(1:6),
Locx1=c(100,121,146,194,162,182),
Locx2=c(148,170,184,236,196,190),
Locy1=c(119,173,104,164,188,142),
Locy2=c(168,180,120,210,190,213))
另外,在上述数据, Locx1
是x的起点和Locx2
是x的端点, Locy1
是y的起点和Locy2
是y的端点。 我想找到的y值在50%(及以上) Locy1
和Locy2
之间是Locx1
和Locx2
在R.我怎么能这样做?
例如,第一行适合这个例子。 y 的起点是Locx1
和Locx2
之间的 (119) 并且 (148-119)/(168-119) 大于 %50。
谢谢
将x
和y
的intersect
s 的lengths
除以完整的y
长度。
## helper FUNs
intl <- function(i) length(i[[1]]:i[[2]]) ## interval length
seq1 <- function(i) i[[1]]:i[[2]] ## seq from `:`
res <- lengths(Map(intersect, apply(y, 1, seq1), apply(x, 1, seq1))) / apply(y, 1, intl)
# [1] 0.6000000 0.0000000 0.0000000 0.3617021 1.0000000 0.1250000
res > .5
# [1] TRUE FALSE FALSE FALSE TRUE FALSE
数据:
d1 <- structure(list(ID = 1:6, Locx1 = c(100, 121, 146, 194, 162, 182
), Locy1 = c(119, 173, 104, 164, 188, 142), Locx2 = c(148, 170,
184, 236, 196, 190), Locy2 = c(168, 180, 120, 210, 190, 213)), class = "data.frame", row.names = c(NA,
-6L))
如果我猜对了
df %>%
filter((pmin(Locy2, Locx2) - pmax(Locy1, Locx1)) / (Locx2 - Locx1) >= 0.5)
ID Locx1 Locx2 Locy1 Locy2
1 1 100 148 119 168
2 6 182 190 142 213
一个有点冗长但希望也更透明的解决方案,它产生与@jay.sf 相同的结果
library(dplyr)
#>
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#>
#> filter, lag
#> The following objects are masked from 'package:base':
#>
#> intersect, setdiff, setequal, union
d1<-data.frame(ID=c(1:6),
Locx1=c(100,121,146,194,162,182),
Locx2=c(148,170,184,236,196,190),
Locy1=c(119,173,104,164,188,142),
Locy2=c(168,180,120,210,190,213))
d2 <- d1 %>% mutate(seqx = purrr::map2(Locx1, Locx2, .f = ~seq(.x, .y, 1)),
seqy = purrr::map2(Locy1, Locy2, .f = ~seq(.x, .y, 1)),
intersection = purrr::map2(seqx, seqy, .f = ~intersect(.x, .y)),
overlap = purrr::map2_dbl(intersection, seqy, .f = ~length(.x)/length(.y)),
my_condition = overlap >= 0.5
)
d2 %>% select(-contains('seq'), -intersection)
#> ID Locx1 Locx2 Locy1 Locy2 overlap my_condition
#> 1 1 100 148 119 168 0.6000000 TRUE
#> 2 2 121 170 173 180 0.0000000 FALSE
#> 3 3 146 184 104 120 0.0000000 FALSE
#> 4 4 194 236 164 210 0.3617021 FALSE
#> 5 5 162 196 188 190 1.0000000 TRUE
#> 6 6 182 190 142 213 0.1250000 FALSE
由 reprex 包 (v0.3.0) 于 2020 年 9 月 3 日创建
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.