Filter values based on other values in r-CodePudding

I have a data frame consisting of several sets that I load into r. I need the entire data frame and then one filtered based on the sets. Currently I bind the sets once to the unfiltered dataframe, then filter the sets and bind them again to the filtered dataframe.

Is there a way to filter the first data frame based on the sets so that I can skip this tedious step?

There are 10 sets, each containing more than 72 hours of time recording every 5 minutes, which start at different times. So I need a solution that does not select the data based on their row number but on the content (here set).

Example:

example<-data.frame(set = c("set1","set1","set1","set1","set1","set1","set1",
                            "set2","set2","set2","set2","set2","set2","set2"),
                    time = c("2022-01-01 05:55", "2022-01-01 06:00", "2022-01-01 06:05", "2022-01-01 06:10", "2022-01-01 06:15", "2022-01-01 06:20", "2022-01-01 06:25",
                             "2022-01-02 05:58", "2022-01-02 06:03", "2022-01-02 06:08", "2022-01-02 06:13", "2022-01-02 06:18", "2022-01-02 06:23", "2022-01-02 06:28")
                    )

I want the time 2022-01-01 06:00 to 2022-01-01 06:15 for set1 and the time 22-01-01 05:58 to 2022-01-02 06:13 for set2.

CodePudding user response：

You could create specific filters for each set like this. With dplyr, you can take advantage of functions like between, first, last, or nth to make it easier.

library(dplyr)
example %>% 
  mutate(time = as.POSIXct(time)) %>% 
  filter((set == "set1" & between(time,
                                as.POSIXct("2022-01-01 06:00"),
                                as.POSIXct("2022-01-01 06:15"))) |
         (set == "set2" & between(time,
                                first(time),
                                as.POSIXct("2022-01-02 06:13"))))

   set                time
1 set1 2022-01-01 06:00:00
2 set1 2022-01-01 06:05:00
3 set1 2022-01-01 06:10:00
4 set1 2022-01-01 06:15:00
5 set2 2022-01-02 05:58:00
6 set2 2022-01-02 06:03:00
7 set2 2022-01-02 06:08:00
8 set2 2022-01-02 06:13:00

CodePudding user response：

If you need to deal with different times for each sets and there are only 10 cases - I would write as an if else statement.

require(data.table)
setDT(example)
res <- rbindlist(lapply(unique(example[, set]), function(i) {
    if (i == "set1") {
        example[time > "2022-01-01 06:00" & time < "2022-01-01 06:15", ]
    } else if (i == "set2") {
        example[time > "2022-01-01 05:58" & time < "2022-01-02 06:13", ]
    # else if (i == "setX") ...
    } else {
        example
    }
}),use.names = TRUE, fill = TRUE)