I love feature caching. It makes for a smooth and flexible development process.
However, there’s one quite painful stone in this shoe.
Whenever I have a workspace reading files from a ZIP, the embedded files are copied to a temporary location, to be read/used from this location.
The problem is that FME always deletes these files as soon as it stops running, also after a partial run, rendering use of feature caching impossible in this case. One always have to start from the beginning.
How about keeping and NOT deleting the created temporary files, until either a new complete run is performed or the workspace/FME is closed ?
It does use a little more resources, but will facilitate the feature caching approach.
Still works as described in 2025.2 beta.
