Speeding up admiral code

[manciniedoardo]

Some of our functions, e.g. the *_joined ones, can be a bit slow, especially with large input data. It may be good to explore ways of speeding them up. Possible options include:

• Using data.table
• Implementing chunking/parallelisation approaches

[bundfussr]

We should also analyse the code on unnecessary steps. When I worked on the memory issue in the *_joined_*() functions I noticed that we spend considerable time on checking on duplicates and that the checks are performed although we know that there are no duplicates.

[bundfussr]

Investigate if using the new vctrs functions like vctr_if_else() improves performance.

[StefanThoma]

The new dplyr version might help as well, just saw this today:
https://tidyverse.org/blog/2026/02/dplyr-performance/