Cutting a continous time series into smaller data frames based on when a condition is met

Question

Ive been trying to clean a and split a data set that contains a timeseries of temperature data. Using R. I've already been able to take this data frame and split it based on other discrete conditions. I am now trying to split it everytime the temperature profile dips near 100 F. This would basically signal the start of the new data frame. I've tried splitting based on time intervals but not all smaller data frames/process last the same. If anyone has an idea on how to tackle this it would be extemele appreciated.

Here is a fragment of the data set. It is quite a bit but its by far the minimum for it to be representative:

structure(list(timestamp = structure(c(1416213315, 1416213317, 
1416213339, 1416213345, 1416213356, 1416213358, 1416213361, 1416213365, 
1416213384, 1416213386, 1416213427, 1416213439, 1416213441, 1416213478, 
1416213490, 1416213494, 1416213524, 1416213557, 1416213560, 1416213595, 
1416213598, 1416213600, 1416213618, 1416213624, 1416213635, 1416213645, 
1416213663, 1416213665, 1416213689, 1416213707, 1416213708, 1416213713, 
1416213715, 1416213719, 1416213729, 1416213732, 1416213742, 1416213744, 
1416213749, 1416213752, 1416213753, 1416213754, 1416213756, 1416213757, 
1416213759, 1416213763, 1416213765, 1416213767, 1416213773, 1416213775, 
1416213777, 1416213784, 1416213788, 1416213790, 1416213792, 1416213794, 
1416213796, 1416213798, 1416213802, 1416213808, 1416213812, 1416213814, 
1416213816, 1416213818, 1416213828, 1416213833, 1416213836, 1416213846, 
1416213847, 1416213849, 1416213854, 1416213856, 1416213858, 1416213863, 
1416213871, 1416213873, 1416213876, 1416213878, 1416213880, 1416213892, 
1416213893, 1416213896, 1416213898, 1416213906, 1416213910, 1416213912, 
1416213914, 1416213916, 1416213917, 1416213921, 1416213925, 1416213927, 
1416213928, 1416213935, 1416213936, 1416213939, 1416213944, 1416213947, 
1416213954, 1416213956), class = c("POSIXct", "POSIXt"), tzone = "CST6CDT"), 
    averagetemp = c(162.6, 162.6, 162.6, 162.6, 162.6, 162.6, 
    162.6, 155.5, 155.5, 152.8, 152.8, 147.1, 147.1, 147.1, 147.1, 
    142.4, 142.4, 139.4, 139.4, 139.4, 139.4, 136.4, 136.4, 135.3, 
    135.3, 135.3, 133, 133, 133, 131.5, 131.5, 124.8, 123.8, 
    123.8, 120.3, 120.3, 118.8, 118.7, 118.7, 145.5, 145.1, 143.9, 
    143.9, 143.9, 134.6, 147.1, 147.1, 147.1, 139.2, 139.2, 144.7, 
    146.4, 135.9, 135.9, 135.9, 137.2, 140.7, 147.2, 137.1, 137.1, 
    137.1, 149.3, 148.2, 148.2, 156.8, 151.4, 150, 150, 158.6, 
    158.6, 162.2, 162.1, 162.1, 160.9, 167.7, 164.5, 162.9, 162.9, 
    171, 165.6, 165.6, 164.1, 164.1, 164.1, 166.3, 166.3, 165.7, 
    164.7, 164.5, 164.5, 164.5, 167.3, 167.3, 167.9, 168, 168, 
    168, 166.2, 166.2, 167.4), fronttemp = c(163.2, 163.2, 159.2, 
    159.2, 159.2, 157.5, 157.5, 157.5, 157.5, 157.5, 157.5, 157.5, 
    157.5, 150.8, 150.8, 150.8, 150.8, 150.8, 150.8, 146.4, 146.4, 
    146.4, 146.4, 146.4, 145.6, 145.6, 145.6, 145.6, 144.9, 144.9, 
    144.9, 144.9, 144.9, 143.9, 143.9, 143.9, 143.9, 143.9, 144.7, 
    144.7, 153.3, 155.5, 157.2, 157.2, 157.2, 157.2, 157.2, 130.7, 
    155.8, 159.9, 159.3, 159.3, 159.3, 174, 172.4, 172.4, 172.4, 
    181.1, 181.1, 181.1, 173, 174.2, 170.3, 172.1, 172.1, 166.5, 
    160, 167, 167, 159.6, 162, 165.9, 167.2, 160.3, 160.3, 163.2, 
    162.5, 162.5, 171.4, 164.7, 164.7, 164.3, 164.3, 164.3, 164.3, 
    170.7, 170.7, 169.1, 169.1, 167.2, 167.2, 167.2, 170.3, 171, 
    170.6, 170.6, 170.6, 170.6, 167, 167), recipe = structure(c(4L, 
    4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 
    4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 
    4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 
    4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 
    4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 
    4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 
    4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L), .Label = c("a", 
    "b", "c", "d", "e", "f", 
    "g", "h", "i", "j"), class = "factor"), 
    recipenumner = c(7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 
    7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 
    7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 
    7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 
    7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 
    7, 7, 7, 7, 7, 7, 7, 7, 7, 7), dischargetemp = c(166.5, 166.5, 
    166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 
    166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 
    166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 
    166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 
    166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 
    166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 
    166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 
    166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 
    166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 
    166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 
    166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5, 166.5), 
    backtemp = c(163.4, 161.8, 161.8, 158.1, 155.8, 155.8, 155, 
    155, 151, 151, 144.8, 144.8, 143.1, 143.1, 137.5, 137.5, 
    135.7, 135.7, 133.5, 133.5, 129.9, 129.9, 129, 129, 129, 
    126.1, 126.1, 124.7, 124.7, 124.7, 123, 111.9, 110.4, 110.4, 
    110.4, 103.7, 102.1, 101.9, 101.9, 141.1, 139.7, 136.2, 136.2, 
    130.1, 130.1, 130.1, 150.6, 150.6, 127.8, 128.9, 135, 135, 
    113.5, 113.5, 113.5, 113.7, 119.6, 119.6, 110.9, 126, 126, 
    132.7, 133.5, 133.5, 146.6, 141.3, 143.4, 143.4, 154.6, 154.6, 
    162.4, 159.5, 159.5, 159.5, 169.9, 165.4, 163.2, 171.4, 171.4, 
    171.4, 165.9, 164, 164.3, 162.3, 162.3, 162.3, 163.1, 161.7, 
    161.7, 161.7, 165.6, 165.6, 161.4, 165.9, 165.9, 166.6, 167.3, 
    165.5, 165.5, 167.7)), .Names = c("timestamp", "averagetemp", 
"fronttemp", "recipe", "recipenumner", "dischargetemp", "backtemp"
), row.names = 168229:168328, class = "data.frame")

This data frame comes from a bigger data frame I was able to split by label names. See code here:

# Step 1: Import the data set

cooker <- read.delim(file.choose(), header=TRUE, "\t")#Choose the tab delimited fileof your choosing
attach(cooker)
cooker$timestamp <- as.POSIXct(timestamp, tz="CST6CDT", "%Y-%m-%d %H:%M:%S")#convert timestamp to POSIXlt to enable time calculations

# Step 2: Subset the data based on recipe/label

cooker_split <- split(cooker, recipe)#split based on recipe
new_names <- names(cooker_split)#created new names based on the names of split
for (i in 1:length(cooker_split)) { #for loop to assign names and creatre individual data frames
  assign(new_names[i], cooker_split[[i]])
}

After I uploaded the data and had it subset into recipe/label names I tried to create subsets based on time intervals like this:

time_interval <- cut(df$timestamp, breaks = "10 sec")
new_df <- cbind(df, time_interval)
batches <- split(new_df, new_df$time_interval)

Issue here is that given the complete data set, the increase and decrease in temp does not last 10 secs, this is only an approximate.

What I want to be able to do, if at all possible is to have R determine that when the temperature has reached 100 or close to 100 F on the variable backtemp, or x condition for backtemp has been met to subset and keep going to subset the next section until the condition canot be met. I sounds like a for-if loop but I cannot even begin to think how it would do that in R. Any help would be appreciated.

Thanks

You need to provide a representative example of your data, and show what you've tried so far. Read this: [how-to-make-a-great-r-reproducible-example](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example/5963610#5963610). — jlhoward, Nov 24 '14 at 16:09
You will find that using `attach` might seems helpful but creates more confusion that it's worth. For instance, your use of `attach` followed by changing 'names' seems like a particularly bad idea. You then create a split object (on an example that has only one level with values for the split criterion) and refer to its 'names' but there are no names of a split object other than the split-criteria names. That split object has a whole bunch of empty dataframes. And then at the end you switch to using `df` rather than any of the objects you used before???? The goals are completely confused. — IRTFM, Nov 24 '14 at 17:11
So given your example at the top (100 rows), how many splits do you expect and where would they occur? — jlhoward, Nov 24 '14 at 19:58

Cutting a continous time series into smaller data frames based on when a condition is met

0 Answers0