如何使用指定的列对将数据框重塑为宽格式

Question

我有以下融化的样本数据框：

df_melted <- data.frame(ID = c(21, 21, 21, 21, 49, 49, 49, 49), instance = c(1, 1, 1, 1, 1, 1, 2, 2), variable = c("causeofdeath", "deathdate", "dob", "gender", "causeofdeath", "deathdate", "deathdate", "causeofdeath"), value_temp = c("BLANK", "MM/DD/YYYY", "BLANK", "F", "BLANK", "BLANK", "MM/DD/YYYY", "BLANK"))

我最终希望使用以下映射文件来生成新值：

df_map <- data.frame(var_name_1 = c("causeofdeath", "causeofdeath", "dob"), val_1 = c("BLANK", "BLANK", "MM/DD/YYYY"), var_name_2 = c("deathdate", "deathdate", "gender"), val_2 = c("MM/DD/YYYY", "BLANK", "F"), new_var_name = c("ida", "idf", "ids"), new_val = c("T60", "T55", "T67"))

如何重塑 df_melted 使其与 df_map 中所示的特定排列相同？，即：

我尝试使用pivot_wider()但无济于事

Answer 1

您可以在添加一些辅助列后尝试

pivot_wider

，例如

p

和

grp

df_melted %>%
    mutate(p = rep(1:2, length.out = n()), grp = ceiling(row_number() / 2)) %>%
    pivot_wider(
        names_from = p,
        values_from = c(variable, value_temp),
        id_cols = c(ID, instance, grp)
    )

你将获得

# A tibble: 4 × 7
     ID instance   grp variable_1   variable_2   value_temp_1 value_temp_2
  <dbl>    <dbl> <dbl> <chr>        <chr>        <chr>        <chr>
1    21        1     1 causeofdeath deathdate    BLANK        MM/DD/YYYY
2    21        1     2 dob          gender       BLANK        F
3    49        1     3 causeofdeath deathdate    BLANK        BLANK
4    49        2     4 deathdate    causeofdeath MM/DD/YYYY   BLANK