CMS: different triggers

[Kjili]

Apparently, triggers are not only different for different data sets but also for different physics objects. This should be investigated further.

[katilp]

There is a predefined set of triggers for each primary dataset as listed, for example, in http://opendata.cern.ch/record/34 under HLT trigger paths. A trigger gives a value of "accepted" for an event if it has passed the trigger criteria at the time of data taking. Each analysis uses one or a combination of these triggers, by selecting for further analysis the events where the trigger of interest has given the value "accepted". So the trigger decision, even if based by the attributes of different physics objects in the event, selects the full event and not part of it (i.e. separate physics objects).

[Kjili]

Okay, does this make the trigger specific to datasets or to measurements (which include the event selection)? As right now the trigger is chosen in combination with a dataset.

[katilp]

The dataset and triggers are specific to a measurement. This is a good point as in most cases the datasets for different measurements are exclusive (in which case the two combinations are equivalent), but it is probably not impossible that the same dataset could be used with different triggers in separate measurement in a single CAP entry.

[Kjili]

Then I will keep the issue as a reminder for now, although the original question is solved and I will check back with this knowledge in mind. Thanks!

[ioannistsanaktsidis]

@Kjili any news on this one ?

[Kjili]

No, but if it helps, we can create a "Pending" milestone for issues that aren't ready for implementation nor deletion as they need more testing/information while at the same time they carry information we might need later or that can be a used for tests.

[Kjili]

@katilp Do you think this is something important to check for the next user tests or should we close it and see if anyone mentions it by himself at some point in time?

[katilp]

This is something we would like to have, and if it is feasible to implement, we should have it.

We need to have a mapping from the primary dataset (Input data filed) name (first part of it, e.g. TauPlusX of /TauPlusX/Run2011A-12Oct2013-v1/AOD) and the possible triggers, and offer users a menu where they can select the triggers.

For Run1 (2010,2011,2012) primary datasets, this trigger mapping is available in https://fwyzard.web.cern.ch/fwyzard/hlt/2010/dataset https://fwyzard.web.cern.ch/fwyzard/hlt/2011/dataset https://fwyzard.web.cern.ch/fwyzard/hlt/2012/dataset

For Run2, I have to check where to get them easiest.

[annatrz]

@katilp any update on this one? We can grab list of triggers from those three files and use for autosuggestions (based on dataset). But as i understood, this list is not complete (?), so we should let user to type something different here as well?

[katilp]

@annatrz The lists are complete for Run1 analyses, so no need for autosuggestions at least to my current knowledge.

[annatrz]

Will be implemented in #1051