How to split/subset a dataframe into multiple dataframes in R - r

Question

I have looked through the web and stackflow and I was not able to find a solution to my problem. I don't know which of dplyr or a loop would be more efficient.
Below an example of a dataframe (my own datasets have more than 10,000 rows) I would like to split in three based on column B (<250) as a list with three objects or as three individual dataframes. Then for each new dataframe, I would like, for example, count the number of points (or length of the dataframe) and the duration (column Time is in seconds). Any suggestion would be really appreciated.
thank you
Martin
dput(mydata)
structure(list(Time = c(1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 0L,
11L, 12L, 13L, 14L, 15L, 16L, 17L, 18L), A = c(4L, 5L, 6L, 7L,
3L, 7L, 8L, 10L, 11L, 8L, 10L, 12L, 14L, 6L, 14L, 16L, 20L, 22L
), B = c(100.25, 150.75, 200, 1000.56, 2000.1, 100, 150, 50,
25.2, 102.25, 152.75, 202, 1002.56, 2002.1, 102, 152, 52, 27.2
)), .Names = c("Time", "A", "B"), class = "data.frame", row.names = c(NA,
-18L))

Answer 1

Grab IRanges from Bioconductor:
runs <- slice(Rle(df$B), upper=250)
This is an RleViews object, with a view (range) for every run under 250. You can extract the width of the views (the number of points that would be in each data frame):
width(runs)
You can split the data frame into a list like this:
blocks <- extractList(df, ranges(runs))
Note that blocks is now a formal SplitDataFrameList.
To compute the duration, you can extract the Time column as an IntegerList and compute the difference between the last and first element of every list element:
time <- blocks[,"Time"]
ptail(time, 1) - phead(time, 1)
This happens without actually forming separate list elements (the list is lazily managed) and so should be fast.

Answer 2

It's not clear how your specifications line up with your expected output. Here are two different method of splitting:
# Gives three groups
split( mydata[mydata$B <250, ] , (1:nrow(mydata[mydata$B <250, ]))%% 3)
# Gives groups of size three
split( mydata[mydata$B <250, ] , (1:nrow(mydata[mydata$B <250, ]))%/% 3)
This shows how to count the numbers of rows from the first method:
> three <- split( mydata[mydata$B <250, ] , (1:nrow(mydata[mydata$B <250, ]))%% 3)
> lapply(three, nrow)
$`0`
[1] 4
$`1`
[1] 5
$`2`
[1] 5