-1

我有一个大致如下所示的数据框（这意味着它是为说明而制作的近似值，而不是您可以通过下面的链接下载或从我粘贴在下面的 dput() 获取的数据框的精确副本）：

March_created_at    March_email March_type  April_created_at April_email    April_type
3/11/12 7:28    jeremy@asynk.ch PushEvent   4/1/12 4:03                     PushEvent
3/11/12 7:28    jeremy@asynk.ch PushEvent   4/1/12 4:03                     PushEvent
3/11/12 7:28    jeremy@asynk.ch PushEvent   4/1/12 4:03                     PushEvent
3/11/12 7:28    jeremy@asynk.ch PushEvent   4/1/12 7:03     high            IssuesEvent
3/11/12 11:06   medium          PushEvent   4/1/12 13:57    medium          PushEvent
3/11/12 11:06   medium          PushEvent   4/1/12 13:57    medium          PushEvent
3/11/12 11:06   medium          PushEvent   4/1/12 13:57    medium          PushEvent
3/11/12 12:46                   PushEvent   
3/11/12 12:46                   PushEvent   
3/11/12 12:46                   PushEvent

完整的数据集可以在这里以 CSV 文件的形式找到

我正在寻找一个接受以下输入的函数：

  1. 一个数据框
  2. 该数据框的某些列
  3. 字符串列表（例如一组电子邮件地址）
  4. 替换字符串（例如“low”）

现在，我希望该函数仅遍历该数据帧的指定列，并将与上面第 3 点中指定的字符串列表不匹配的所有字符串（以及空单元格）替换为第 4 点中的替换字符串。但是，只有在以下条件成立时才应该这样做：

所考虑的单元格需要具有同一月份的时间戳。

例如，假设我们要替换“March_email”列中第 8 行的空单元格。我可以看到在“March_created_at”列的第 8 行有一个时间戳，所以我可以继续用指定的字符串（例如“low”）替换这个空单元格。但是，请查看“April_email”列中的第 8 行。此单元格也是空的，“April_created_at”列中第 8 行的单元格也是空的。在这种情况下，什么都不应该做（即没有插入字符串）。

我想这样做的原因是某些单元格只是空的，因为没有数据，所以不应该插入任何内容。其他单元格是空的，因为数据丢失了，所以我需要根据我上面指定的函数来估算数据。

我怎样才能在 R 中做到这一点？

附录：这是数据集头部的 dput()：

structure(list(March_created_at = c("2012-03-11 07:28:04", "2012-03-11 07:28:04", 
"2012-03-11 07:28:04", "2012-03-11 07:28:19", "2012-03-11 07:28:19", 
"2012-03-11 07:28:19"), March_actor_attributes_email = c("jeremy@asynk.ch", 
"jeremy@asynk.ch", "jeremy@asynk.ch", "jeremy@asynk.ch", "jeremy@asynk.ch", 
"jeremy@asynk.ch"), March_type = c("PushEvent", "PushEvent", 
"PushEvent", "PushEvent", "PushEvent", "PushEvent"), April_created_at = c("2012-04-01     04:03:13", 
"2012-04-01 04:03:13", "2012-04-01 04:03:13", "2012-04-01 07:03:11", 
"2012-04-01 07:03:11", "2012-04-01 07:03:11"), April_actor_attributes_email = c("", 
"", "", "high", "high", "high"), April_type = c("PushEvent", 
"PushEvent", "PushEvent", "IssuesEvent", "IssuesEvent", "IssuesEvent"
), May_created_at = c("2012-05-01 00:16:05", "2012-05-01 00:16:05", 
"2012-05-01 00:16:05", "2012-05-01 01:03:19", "2012-05-01 01:03:19", 
"2012-05-01 01:03:19"), May_actor_attributes_email = c("john.firebaugh@gmail.com", 
"john.firebaugh@gmail.com", "john.firebaugh@gmail.com", "mitch.tishmack@gmail.com", 
"mitch.tishmack@gmail.com", "mitch.tishmack@gmail.com"), May_type = c("PushEvent", 
"PushEvent", "PushEvent", "IssueCommentEvent", "IssueCommentEvent", 
"IssueCommentEvent"), June_created_at = c("2012-06-01 00:25:05", 
"2012-06-01 00:25:05", "2012-06-01 00:25:05", "2012-06-01 00:42:29", 
"2012-06-01 00:42:29", "2012-06-01 00:42:29"), June_actor_attributes_email =     c("michaelklishin@me.com", 
"michaelklishin@me.com", "michaelklishin@me.com", "", "", ""), 
    June_type = c("IssueCommentEvent", "IssueCommentEvent", "IssueCommentEvent", 
    "PushEvent", "PushEvent", "PushEvent"), July_created_at = c("2012-07-01 13:46:20", 
    "2012-07-01 13:46:20", "2012-07-02 11:53:37", "2012-07-02 11:53:37", 
    "2012-07-02 12:27:30", "2012-07-02 12:27:30"), July_actor_attributes_email = c("medium", 
    "medium", "ryoqun@gmail.com", "ryoqun@gmail.com", "ryoqun@gmail.com", 
    "ryoqun@gmail.com"), July_type = c("PushEvent", "PushEvent", 
    "CreateEvent", "CreateEvent", "PushEvent", "PushEvent"), 
    August_created_at = c("2012-08-01 00:04:09", "2012-08-01 00:04:09", 
    "2012-08-01 00:04:42", "2012-08-01 00:04:42", "2012-08-01 00:05:04", 
    "2012-08-01 00:05:04"), August_actor_attributes_email = c("jeremy@asynk.ch", 
    "jeremy@asynk.ch", "jeremy@asynk.ch", "jeremy@asynk.ch", 
    "jeremy@asynk.ch", "jeremy@asynk.ch"), August_type = c("IssueCommentEvent", 
    "IssueCommentEvent", "IssuesEvent", "IssuesEvent", "IssueCommentEvent", 
    "IssueCommentEvent"), September_created_at = c("2012-09-01 18:12:24", 
    "2012-09-01 18:12:24", "2012-09-01 23:51:18", "2012-09-01 23:51:18", 
    "2012-09-02 00:34:54", "2012-09-02 00:34:54"), September_actor_attributes_email = c("ryoqun@gmail.com", 
    "ryoqun@gmail.com", "ryoqun@gmail.com", "ryoqun@gmail.com", 
    "ryoqun@gmail.com", "ryoqun@gmail.com"), September_type = c("CommitCommentEvent", 
    "CommitCommentEvent", "CreateEvent", "CreateEvent", "PushEvent", 
    "PushEvent"), October_created_at = c("2012-10-01 07:48:38", 
    "2012-10-01 10:01:40", "2012-10-01 10:01:43", "2012-10-01 10:17:00", 
    "2012-10-01 16:08:29", "2012-10-01 18:06:46"), October_actor_attributes_email = c("medium", 
    "medium", "medium", "medium", "", "core"), October_type = c("PushEvent", 
    "IssuesEvent", "PushEvent", "PushEvent", "ForkEvent", "PullRequestEvent"
    )), .Names = c("March_created_at", "March_actor_attributes_email", 
"March_type", "April_created_at", "April_actor_attributes_email", 
"April_type", "May_created_at", "May_actor_attributes_email", 
"May_type", "June_created_at", "June_actor_attributes_email", 
"June_type", "July_created_at", "July_actor_attributes_email", 
"July_type", "August_created_at", "August_actor_attributes_email", 
"August_type", "September_created_at", "September_actor_attributes_email", 
"September_type", "October_created_at", "October_actor_attributes_email", 
"October_type"), row.names = c(NA, 6L), class = "data.frame")
4

1 回答 1

3

像这样的东西怎么样：

myfun <- function(month, DF, matches, replacement) {
  email.col <- paste0(month, '_actor_attributes_email')
  date.col  <- paste0(month, '_created_at')


  DF[[email.col]] <- ifelse(DF[[date.col]] != '' & !DF[[email.col]] %in% matches,
                            DF[[email.col]],
                            replacement)

  return (DF[, c(date.col, email.col)])
}

myfun(dat, 'April', 'high', 'foo')

#          April_created_at April_actor_attributes_email
# 1 2012-04-01     04:03:13                          foo   
# 2     2012-04-01 04:03:13                          foo   
# 3     2012-04-01 04:03:13                          foo   
# 4     2012-04-01 07:03:11                          high
# 5     2012-04-01 07:03:11                          high
# 6     2012-04-01 07:03:11                          high

然后，你可以喂它几个月...

out <- lapply(list('March', 'April', 'May'), 
              myfun, DF=dat, matches='', replacement='foo')

您可以快速将其恢复到 data.frame 中。和plyr

as.data.frame(unlist(out, recursive=FALSE))

还有很多其他的方法和选择，但这应该会给你一个很好的开始。

于 2012-10-18T22:33:57.790 回答