QA Scalars

Peter Jacobs, July 11, 2000

Overview of scalars

QA scalars are generated by several of the QA macros and are made available for a given data set in autoQA via the "QA details" and "Compare to Reference" buttons. As of September '00, the scalars are presented by autoQA within restricted multiplicity classes for the real data, with the classes having the same meaning. "LM" means global track multiplicity between 50 and 500, "MM" between 500 and 2500, and "HM" above 2500. No multiplcity classes are defined for MC data. The scalars for a given macro and multiplicity class are displayed by pushing the appropriately labelled button.

bfcread_dstBranch

This macro displays quantities from the tables-based dst (i.e. not StEvent). All scalars are for the first event in the run, except for those labelled "avg_XXX_per_event", which are averaged over all events in the file. The scalars from this macro are useful to look at for the MC nightly tests to test for software stability, because every event in the file has identical underlying physics and the same MC events are used each night. The scalars from this macro have much less meaning when looking at real data, because the first event in a real data file could have a potentially wide range of properties. For real data, it is useful to look at these scalars only to verify that some reasonable values are present - the specific values cannot be specified run to run.

bfcread_eventBranch

The scalars from this macro are by far the most important for the shift crews to look at. The quantities are extracted from StEvent. The mean, min, max and rms are for a given quantity are calculated over all events in the run passing the multiplciity cut (if appropriate). The names of the scalars are self-explanatory and are valid for both MC and real data.

bfcread_geantBranch

Only for MC data. Shows GSTAR related quantities for the first event in the file.

bfcread_runcoBranch

These scalars show the production "run conditions", which are the same for all events in the file. Their specific values for a given run are best specified by the production crew, whom you should consult in case of questions. Anomolies in these scalars would presumably manifest themselves in other features of the analysis (e.g. certain claibrations not run when they should have been), so monitoring of these scalars is not essential.

bfcread_tagsBranch

These scalars show quantities from the tag database, averaged over events in the run passing a given multiplicity cut. Most of these scalars are not useful for QA purposes and the shift crew can ignore them.