What is the difference between <schema-flattening>...
# getting-started
r
What is the difference between schema-flattening and record-flattening capabilities? Looking at the SDK code, it seems
schema-flattening
controls the application of flattening config, schema flattening and record flattening (i.e. does more than the capability name implies).
record-flattening
does not seem to be respected.
schema-flattening
is also missing from the SDK docs.
e
What is the difference between schema-flattening and record-flattening capabilities?
Hmm. This went under my radar 😬. It seems they were added at different points: • `record-flattening`: https://github.com/meltano/sdk/commit/fd95f124ccdaafc6e190b31894c91b8adaf1d01b • `schema-flattening`: https://github.com/meltano/sdk/commit/010b41eb80a00cd06251e5976e05d666887982f3 I forgot the rationale, if there was any, for unfolding them like that. I see AJ titled the feature
Schema Flattening Applications
when he implemented it and MeltanoHub uses
schema-flattening
.
record-flattening
does not seem to be respected
can you expand on what you mean by this?
schema-flattening
is also missing from the SDK docs.
We can add it after we clarify this mess 😅
r
can you expand on what you mean by this?
record-flattening
is a defined target capability, but I can't see where it is used - if at all.
Feels like the capability (in an ideal world) should just be
flattening
since the combination of
schema-flattening
capability and flattening config operate on both schema and records.
e
I agree. And the behavior is the same in both taps and targets.
r