Question

我的问题类似于this question。我试图tidyr::gather多列。但是，链接中提供的解决方案不太理想，因为所有列的属性通常不相同，因此它们被删除。

注意，我知道如何使用base R执行此操作，但我正在尝试学习如何使用tidyr和/或dplyr执行等效操作。

下面我模拟了一些数据（很差但很快），这些数据说明了我经常发现的情况（尽管我通常有更多的列遵循这种相同的模式）。我已经为stats::reshape提供了基本解决方案，因此您可以看到我想要的输出。

非常感谢任何帮助。

set.seed(123)
male_g6 <- rbinom(100, 1, .5)
ell_g6 <- rbinom(100, 1, .1)
sped_g6 <- rbinom(100, 1, .15)
pullouts_g6 <- rbinom(100, 5, .1)
disability_g6 <- replicate(100, 
                sample(
                    c("asd", "cd", "ed", "hi", "id", "ohi", "ld", "none"),
                    1,
                    prob = c(rep(0.01, 6), 0.05, 0.89)
                    )
                 )
score_g6 <- rnorm(100, 200, 10)
score_g7 <- score_g6 + 5 + rnorm(100, 0, 2)
score_g8 <- score_g7 + 5 + rnorm(100, 0, 2)

d <- data.frame(
        SID = 1:100,
        male_g6 = male_g6,
        male_g7 = male_g6,
        male_g8 = male_g6,
        ell_g6 = ell_g6,
        ell_g7 = ell_g6,
        ell_g8 = ell_g6,
        sped_g6 = sped_g6,
        sped_g7 = sped_g6,
        sped_g8 = sped_g6,
        pullouts_g6 = pullouts_g6,
        pullouts_g7 = pullouts_g6,
        pullouts_g8 = pullouts_g6,
        disability_g6 = disability_g6,
        disability_g7 = disability_g6,
        disability_g8 = disability_g6,
        score_g6 = score_g6,
        score_g7 = score_g7,
        score_g8 = score_g8
    )

使用base reshape

ld <- stats::reshape(d,
        idvar = "SID",
        varying = list(
            c("male_g6", "male_g7", "male_g8"),
            c("ell_g6", "ell_g7", "ell_g8"),
            c("sped_g6", "sped_g7", "sped_g8"),
            c("pullouts_g6", "pullouts_g7", "pullouts_g8"),
            c("disability_g6", "disability_g7", "disability_g8"),
            c("score_g6", "score_g7", "score_g8")
            ),
        v.names = c("male", "ell", "sped", "pullouts", "disability", "score"),
        times = 6:8,
        timevar = "Grade",
        direction = "long"
    )
ld <- ld[order(ld$SID), ]

Answer 1

您需要收集超出您想要结束的内容，以便您可以将成绩级别与标题分开，然后您可以将其扩展回广泛的形式：

ld2 <- d %>% gather(var, val, -SID) %>%     # gather to long form
    # separate grade from variable names
    separate(var, c('var', 'grade'), sep = '_g', convert = TRUE) %>% 
    spread(var, val, convert = TRUE)    # spread back to wide

head(ld2)

##   SID grade disability ell male pullouts    score sped
## 1   1     6         cd   0    0        1 196.2440    0
## 2   1     7         cd   0    0        1 203.2739    0
## 3   1     8         cd   0    0        1 211.1347    0
## 4   2     6       none   0    1        0 194.3812    1
## 5   2     7       none   0    1        0 195.3957    1
## 6   2     8       none   0    1        0 202.4890    1

tidyr ::收集多个不同类型的列

使用base reshape

1 个答案: