tdiff | tree differencing library for javascript

I am looking for help with creating and tracking user sessions and activities within sessions using R. At a high level I have a column of user Ids and a column of timestamps.

For each user ID I want to calculate the time difference between timestamps and use that to assign a session number based on a time limit value. (i.e. the first timestamp for the user gets a session number of 1. If the next time stamp is < 30 minutes later, the session number would remain as 1, otherwise it would be incremented to 2, and so on.)

Once a session number is assigned for each user action observation I then want to sequence the activities within each session. The first observation for each session would have an activity number of 1, the second observation within the session would have an activity number of 2 and so on. When the session number changes (or the user ID changes) the activity number would reset back to 1.

I have the following repro which does what I am looking for however, this code is very slow on my full dataset (1.4 million rows). In particular, the calls to difftime and the for loops are very slow. I am wondering if there is a better way to do this and am open to multiple ideas. Perhaps there is a way to do this without for loops using purrr or perhaps this is an excellent use case for a parallel library. Happy to hear any suggestions for how to make this process easier.

I wish to extract all columns for rows 4, 11 and so on. If you look at my posted data, my wish is to extract row values that are present immediately before an 'A' in column 'xsampa'. For example, all the columns for row 4 (that occur before row 5 that contains an 'A' in column 'xsampa'). I can manually extract them but anything better will definitely save me some labour.

Many thanks if you help me out.

I found that the gettickdiff64 function sometimes results in 18446744073709551600 (or 18446744073709551601) and causes the program to run incorrectly. Normally does not have a result greater than 300000

what might this be about? Should I always do extra checks against this problem?

it is 32 bit VCL application. I use Delphi 10.4.1( its indy version 10.6.2.0 ) Running on: 64 bit Windows Server 2012 R2 Foundation / intel xeon cpu E3-1225 v5 3.3 Ghz.

The code structure is as follows:

This is how the code currently looks like.

Firstly, I am fairly new to python/ML in general. I am attempting to utilize the model depicted at stackabuse over my own data set.

Everything flows smoothly until I get ready to run the epochs.

In debugging I see that it is failing on CrossEntropyLoss function and I get the error expected long found double. The data set it appears to fail on is the my tdiff column that I calculated but I can't seem to figure out how to convert it to a long.

Is there something that I'm missing in trying to figure this out?

To be clear, this is what I THINK it is based on my extremely limited knowledge on the function:

Say I am in initial user namespace and there is an empty root-owned file in some directory:

A sample picture attached The sample dataset file I have a dataset (please see the attached file), in which I wish to sum the numeric column 'tdiff' based on a specific criteria, e.g. row (1 + 2), row (3 + 4) but not row (11,12,13,14). I have tried these but no luck,

I've been testing java reflection and I found that, when using execute(), the time needed to finish the method varies a lot.

Here is the code that I'm using:

Whish to have time duration/accumulation of time diff as long as "state" == 1 is active and else 'off'