In ApacheBeam once you have some PCollection input you can do
input.aplly(new ParDo())
however BigQueryIO.read() can be applied only on the Pipeline instance, so my question is how can I make BigQueryIO.read() wait till some other DoFn finishes or produces at least 1 output, should it be a different pipeline where I'll put BigQueryIO or can it be done within the same one?
I don't think it's possible to make BigQueryIO.read() wait for some input since, actually, it creates a PTransform<PBegin, PCollection<T>> where PBegin input type says that it's supposed to be executed in the beginning of your pipeline.
I also don't see any other "read" PTransform's implemented in BigQueryIO connector that would accept any input PCollection.
So, very likely it will be easier run it as a different pipeline and use something like Apache Airflow to orchestrate them.