我有一个具有多个列和行的数据框（df），例如：
A B C
0.6 a. b
0.9 c. d
1.1. e. f
1.2 g. h
1.4 I l
1.5. m. n
5.0 o. p
5.3 q. r
5.6. s. t
6.1. u v
6.5. w. z
6.9. y a
7.0. b. c
我正在寻找的代码应该计算A列中每个连续值之间的差异（0.9-0.3 = 0.3、1.1-0.9 = 0.2等），如果差异较大，则应确定一个阈值（在此设置）等于3，但可以不同），它将在该间隙之前和之后分配一定数量的行（在这种情况下，假设为3，但是也可以不同），其中该间隙大于阈值设置。 因此，在这种情况下，将保留5.0-1.5 = 3.5，该值大于3、1.5之前的3行和5.0之后的3行，其余部分将被删除。 对如何编写这样的代码有任何想法吗？
输出：
A B C
1.1. e. f
1.2 g. h
1.4 I l
1.5. m. n
5.0 o. p
5.3 q. r
5.6. s. t
6.1. u v
我有多个数据帧，所以A列中的值不同，代码应逐个查看每个数据帧，并根据阈值设置找出A列中的间隙在哪里。
dput格式的数据。
输入：data.frame df1。
df1 <-
structure(list(A = c(0.6, 0.9, 1.1, 1.2, 1.4,
1.5, 4, 4.3, 4.6, 5.1, 5.5, 5.9, 6),
B = structure(c(1L, 3L, 4L, 5L, 6L, 7L, 8L,
9L, 10L, 11L, 12L, 13L, 2L), .Label = c("a.",
"b.", "c.", "e.", "g.", "I", "m.", "o.",
"q.", "s.", "u", "w.", "y"), class = "factor"),
C = structure(c(2L, 4L, 5L, 6L, 7L, 8L, 9L,
10L, 11L, 12L, 13L, 1L, 3L), .Label = c("a",
"b", "c", "d", "f", "h", "l", "n", "p",
"r", "t", "v", "z"), class = "factor")),
row.names = c(NA, -13L), class = "data.frame")
输出：data.frame out。
out <-
structure(list(A = c(1.1, 1.2, 1.4, 1.5, 4,
4.3, 4.6, 5.1), B = structure(1:8,
.Label = c("e.", "g.", "I", "m.", "o.",
"q.", "s.", "u"), class = "factor"),
C = structure(1:8, .Label = c("f", "h", "l",
"n", "p", "r", "t", "v"), class = "factor")),
row.names = c(NA, -8L), class = "data.frame")
这是我的df：
structure(list(POS = c(207687374L, 207689227L, 207690871L, 207691563L,
207693563L, 207694165L, 207694357L, 207738077L, 207739127L, 207740272L,
207740868L, 207747296L, 207747984L, 207748107L), SNP = c("rs12130494",
"rs4844601", "rs10863358", "rs77357299", "rs12043913", "rs61822967",
"rs11117991", "rs7515905", "rs3886100", "rs12038575", "rs34883952",
"rs1752684", "rs17046851", "rs10127904"), Std_iHS = c(-1.52176,
-1.51905, -1.50286, 0.656487, -1.45251, 0.84325, -1.06089, -1.41041,
1.29513, 1.21325, 0.456717, -1.00933, -1.71468, 0.265969)), row.names =
21:34, class = "data.frame")
输出：
structure(list(POS = c(207691563L,
207693563L, 207694165L, 207694357L, 207738077L, 207739127L, 207740272L,
207740868L, ), SNP = c( "rs77357299", "rs12043913", "rs61822967",
"rs11117991", "rs7515905", "rs3886100", "rs12038575", "rs34883952",
), Std_iHS = c( 0.656487, -1.45251, 0.84325, -1.06089, -1.41041,
1.29513, 1.21325, 0.456717, )), row.names = 21:34, class = "data.frame")
答案 0 :(得分：2)
使用基数R，您可以执行以下操作：
limit = 2
df1[match(unique(c(sapply(which(diff(df1$A)>limit),function(x)(x-3):(x+4)))),1:nrow(df1)),]
A B C
3 1.1 e. f
4 1.2 g. h
5 1.4 I l
6 1.5 m. n
7 4.0 o. p
8 4.3 q. r
9 4.6 s. t
10 5.1 u v
答案 1 :(得分：1)
您的示例数据框似乎没有超过3.0的任何跳转，但是此代码应该可以工作：
limit <- 2.0
structure(list(A = c(0.6, 0.9, 1.1, 1.2, 1.4,
1.5, 4, 4.3, 4.6, 5.1, 5.5, 5.9, 6),
B = structure(c(1L, 3L, 4L, 5L, 6L, 7L, 8L,
9L, 10L, 11L, 12L, 13L, 2L), .Label = c("a.",
"b.", "c.", "e.", "g.", "I", "m.", "o.",
"q.", "s.", "u", "w.", "y"), class = "factor"),
C = structure(c(2L, 4L, 5L, 6L, 7L, 8L, 9L,
10L, 11L, 12L, 13L, 1L, 3L), .Label = c("a",
"b", "c", "d", "f", "h", "l", "n", "p",
"r", "t", "v", "z"), class = "factor")),
row.names = c(NA, -13L), class = "data.frame") %>%
mutate(diffA = A - lag(A, 1)) %>%
mutate(over_limit = diffA > limit) %>%
mutate(before_limit = lag(over_limit, 1) | lag(over_limit, 2),
after_limit = lead(over_limit, 1) | lead(over_limit, 2)) %>%
rowwise() %>%
mutate(subset_filter = any(over_limit, after_limit, before_limit)) %>%
ungroup() %>%
filter(subset_filter) %>%
select(-c(subset_filter, diffA, over_limit, before_limit, after_limit))
dput（）格式的输出：
structure(list(A = c(1.4, 1.5, 4, 4.3, 4.6),
B = structure(6:10, .Label = c("a.", "b.", "c.", "e.", "g.", "I", "m.", "o.", "q.", "s.", "u", "w.", "y"), class = "factor"),
C = structure(7:11, .Label = c("a", "b", "c", "d", "f", "h", "l", "n", "p", "r", "t", "v", "z"), class = "factor")),
class = c("tbl_df", "tbl", "data.frame"),
row.names = c(NA, -5L), .Names = c("A", "B", "C"))