[英]From long to wide formats just based on two columns Rstudio
这是我的数据框:
我有一个六列的数据框,最后一列包含值。 列“代码”包括s和d。 “性别”一栏包括M和F。我的后代中有2000个后代。
seq parent code Sex offspring Value
1 49032 s M J44010_CCG7YANXX_2_661_X4 -0.38455056
2 48741 s M J44010_CCG7YANXX_2_661_X4 0.10574340
3 48757 s M J44010_CCG7YANXX_2_661_X4 0.39572906
4 48465 d f J44010_CCG7YANXX_2_661_X4 0.43409006
5 48521 d f J44010_CCG7YANXX_2_661_X4 0.40337447
6 48703 d f J44010_CCG7YANXX_2_661_X4 -0.38148980
父列中包含男性和女性的ID。 我想将雄性/雌性旁边的雌性/雌性ID,雌性/雌性代码和雌性/雌性列为一列,并将雄性值和雌性值分开保持。 因此,“价值”将分为两部分。
数据框如下所示:
'seq''parent1''sirecode''Sex''parent2''damcode''Sex''offspring''sireValue' 'damvalue'
1 49032 s M 48465 d f J44010 -0.38455056 0.43409006
2 48741 s M 48521 d f J44010 0.10574340 0.40337447
3 48757 s M 48703 d f J44010 0.39572906 -0.38148980
因此,每个后代将有3或4对父母。
我试图在上面使用dcast
函数。
我们可以在创建序列列之后使用dcast
library(data.table)
setDT(df1)[, n := seq_len(.N), .(code, Sex)]
dcast(df1, n + offspring ~ rowid(n), value.var = c('parent', 'code', 'Sex', 'Value'), sep = "")
# n offspring parent1 parent2 code1 code2 Sex1 Sex2 Value1 Value2
#1: 1 J44010_CCG7YANXX_2_661_X4 49032 48465 s d M f -0.3845506 0.4340901
#2: 2 J44010_CCG7YANXX_2_661_X4 48741 48521 s d M f 0.1057434 0.4033745
#3: 3 J44010_CCG7YANXX_2_661_X4 48757 48703 s d M f 0.3957291 -0.3814898
在base R
,我们可以使用reshape
df1$n <- with(df1, ave(seq_along(Sex), Sex, FUN = seq_along))
df1$n1 <- with(df1, ave(n, n, FUN = seq_along))
reshape(df1[-1], idvar = c('n', 'offspring'), timevar = 'n1', direction = 'wide' )
df1 <- structure(list(seq = 1:6, parent = c(49032L, 48741L, 48757L,
48465L, 48521L, 48703L), code = c("s", "s", "s", "d", "d", "d"
), Sex = c("M", "M", "M", "f", "f", "f"),
offspring = c("J44010_CCG7YANXX_2_661_X4",
"J44010_CCG7YANXX_2_661_X4", "J44010_CCG7YANXX_2_661_X4",
"J44010_CCG7YANXX_2_661_X4",
"J44010_CCG7YANXX_2_661_X4", "J44010_CCG7YANXX_2_661_X4"),
Value = c(-0.38455056,
0.1057434, 0.39572906, 0.43409006, 0.40337447, -0.3814898)),
class = "data.frame", row.names = c(NA, -6L))
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.