When it breaks

Organised by what you saw, not by what is wrong

Not yet written. The structure below is settled; the prose is not.

Each entry states what you saw, why it happens, the fix, and whether the fix is yours or your admin’s.

Entries to write

  • Polygons come back truncated or unreadable
  • unknown type 0
  • A package upgrade appears not to have taken
  • install.packages("sf") fails at configure
  • terra fails to load
  • A BINARY value arrives far longer than it should be
  • columns = list(...) rejected
  • context = rejected
  • Hashes differ between your session and a worker
  • An R notebook refuses to attach to a cluster
  • A spark_apply() job is slower than the serial version
  • spark_apply(group_by = ...) raises a deprecation warning from pandas2ri.activate()

Next

If nothing above matches what you saw, the cause is often vocabulary rather than a fault: Words translates the term in the error message. If it looks like a missing package, see Packages.

This page rests on the same claims as the pages that link to it, keyed to the symptom rather than the mechanism.