Thanks Biao, your insights were very helpful for me! I finally could debug properly, a couple of things were not correct there, in fact. Also, I think I'm not modeling properly the behaviour of the pattern so I definitely going to investigate better this.
Regards *Ana Gómez González* <http://twitter.com/angoglez> <https://www.linkedin.com/in/angoglez/> El mié, 19 abr 2023 a las 8:26, Biao Geng (<biaoge...@gmail.com>) escribió: > Hi Ana, > Thanks for the codes. I just want to share my own experience when > debugging CEP patterns: > 1. It should work when adding breakpoints in IntelliJ. But you should add > the breakpoint on the filter function like the call of isBetweenHoursRange > or the line65 in your gist. The reason is that the pattern object will be > built once, when no events entered as the env.execute is not called. But > the filter method in condition object of the pattern will be executed to > detect if the NFA should advance to next phase. > 2. You can add some System.out.println in the condition so you may know > which condition is matched and which condition is not matched. It is > helpful in most cases. > > For test writing, you can reference to codes > of org.apache.flink.cep.CEPITCase. > My final thought is that your line63 and 64 use 2 where clauses so these > 2 condition will be joined with "and"(i.e. for a single event, it should > fulfill both of the 2 conditions). You can check if that is what you really > want. > > Best, > Biao Geng > > Ana Gómez González <angog...@gmail.com> 于2023年4月19日周三 05:09写道: > >> Here's a link for the previous code in a gist so you don't struggle with >> the format. Sorry. >> >> https://gist.github.com/angoglez/d9eb6e12f259aba306387b5c23488fb4 >> >>>