我在R中的过滤选项中挣扎。我有这个数据集：

patient_id    period     TREAT_CAT               Outcome
1            -3228 days pre-treatment                Pink
1            -3170 days pre-treatment                Pink
1              100 days post-treatment               Blue
1              200 days post-treatment               Pink
2            -2900 days pre-treatment                Blue
2                0 days post-treatment               Pink
2              100 days post-treatment               Pink

structure(list(patient_id = c(1, 1, 1, 1, 2, 2, 2), 
    period = structure(c(-3228, -3170, 100, 200, -2900, 
    0, 100), class = "difftime", units = "days"), 
    TREAT_CAT = structure(c( 1L, 1L, 2L, 2L, 1L, 2L, 2L), levels = c("pre-treatment", "post-treatment"), class = "factor"), 
    Outcome = c("Pink", "Pink", "Blue", "Pink", 
    "Blue", "Pink", "Pink"), row.names = c(19L, 26L, 24L, 3L, 7L, 29L, 20L), class = "data.frame")

我想过滤治疗前组最接近0的“周期”和治疗后组最接近0的“周期”。
我试过这样的方法

df2 <- df %>%
  group_by(patient_id) %>% 
  filter((TREAT_CAT=="pre-treatment" & period == min(period)) | (TREAT_CAT=="post-treatment" & period == max(period))) %>% 
  filter(n() == 2)

但很明显，它给了我两个时间段从0开始的最大值，我也尝试了两个组的max（时间段），但它不起作用，因为max（时间段）只发生在治疗后组，导致0变量。
我希望是这样的

patient_id    period     TREAT_CAT               Outcome
1            -3170 days pre-treatment                Pink
1              200 days post-treatment               Pink
2            -2900 days pre-treatment                Blue
2              100 days post-treatment               Pink

你能帮帮忙吗？
先谢了