GitHub user dave-csc added a comment to the discussion: Dynamic creation of stream fields
> Hi! Thanks for your suggestion! Is the term "schema" in this context > referring to the set of the stream's field-names together with their > datatypes and other meta information? Well... yes. In Hop this is actually called "metadata" (sorry, I'm still a little biased from T***nd '^_^) > Let's assume we have one datetime1 field. So the idea would be to generate > first a stream containing the new fieldnames and datatypes (e.g.: > fk_datetime1_date / integer, fk_datetime1_time / integer) out of the existing > ones via the "Metadata structure" transform, and then use this output stream > as Input to the Metadata Injection? If I got correctly, you need a way to calculate some field names from other field names: in a pipeline you usually transform values into other values instead. The **Metadata structure of stream** lets you convert the field names, types and other info into data, and so you can use them for processing (by using other transforms). But to use this as a schema/metadata again, you need to inject this result into another pipeline (which can still have a "real" data flow as input, and the calculated metadata supporting along) GitHub link: https://github.com/apache/hop/discussions/5005#discussioncomment-12463186 ---- This is an automatically sent email for [email protected]. To unsubscribe, please send an email to: [email protected]
