Orchestration

What is incremental computation?

Incremental computation recomputes only the rows and downstream columns a change actually affects, instead of rerunning the whole pipeline.

Updated

How it works

  • Dependencies between columns are known.
  • An edit marks the descendants of that row.
  • Everything else stays cached.

What it is not

It is not a nightly full refresh, and it is not a cache you remember to invalidate.

incremental computation: this, and the thing it is confused with

incremental computation: this, and the thing it is confused with
ThisNot this
WorkChanged rowsThe whole table
InvalidationRecorded dependenciesHope
CostModel calls for the deltaModel calls for the corpus

Where Pixeltable fits

Pixeltable tracks those dependencies on computed columns. A new file recomputes that row’s descendants.

Questions

How does incremental computation work?
Dependencies between columns are known. An edit marks the descendants of that row. Everything else stays cached.
What is incremental computation often confused with?
It is not a nightly full refresh, and it is not a cache you remember to invalidate.