bigdata - Can BDD work for Big Data ETL testing? -

i wandering if uses bdd testing big data etl application? can see how bdd can used testing applications having client interact them, in case of big data etl application there no client interaction hard see 'when' might use. example: give 100 event of type occur , 50 event of type b occur after 5 minute database rows should be: |type|count|bucket| |a|100|1| |b|50|2|

but seems wrong. 1 insight?

can give me example of you'd expect see in etl output?

there couple of responses give this. 1 might different kinds of database rows you'd expect, , fact of them repeated, not others. struck me weird, if you're used working star schemas you'll notice other differences instead.

normally i'd steer people away talking database, if you're working star schemas, think it's ok mention facts , dimensions (i haven't worked etl lot, remember talking through specific examples of these , expect see).

the alternative use client.

i saw said there no client; however, there's always client, if it's 1 might exist in future. there implications etl run across security, performance , access, amongst others. it's worth having client, if it's string-based or sql-based toy, explore things might trip up.

why doing this? what's new thing business or users or customers able when in place, can't already? , can hold of example of that?

"we'll able understand how x performing against y standard."

great. can give me example of x, y, , standard? how measure performance? data looking for? should able see data? can think of scenario shouldn't able access that?

those examples become etl equivalent of scenarios; conversations retain same pattern. end automating them @ different level, since api machine-oriented rather human-oriented, , of conversations monitoring instead of testing. conversations should still people.

your "when" query or report run, within data, permission , security context in run it.