Large Data With Dask Skill

Iron Laws

ALWAYS call dask.compute() only once at the end of a pipeline — multiple intermediate compute() calls break the lazy evaluation graph and eliminate Dask's ability to fuse and parallelize operations.
NEVER use df.apply(lambda ...) with Dask DataFrames for element-wise operations — Pandas-style apply forces row-by-row Python execution that bypasses Dask's vectorized C extensions and is slower than single-threaded Pandas.
ALWAYS specify partition sizes explicitly when reading large datasets (blocksize= for CSV, chunksize= for Parquet) — auto-detected partition sizes frequently produce thousands of tiny partitions (slow scheduler overhead) or a single giant partition (no parallelism).
NEVER call len(df) or df.shape on a Dask DataFrame without wrapping in compute() — these trigger immediate full dataset computation and negate lazy evaluation.
ALWAYS use dask.distributed.Client for multi-machine or CPU-bound workloads — the default threaded scheduler serializes Python-heavy operations due to the GIL; the distributed scheduler bypasses this.

Anti-Pattern	Why It Fails	Correct Approach
Multiple `compute()` calls in pipeline	Breaks lazy graph; forces data to materialize and re-partition at each call	Build complete computation graph first; call `compute()` once at the end
`df.apply(lambda ...)` on large DataFrames	Row-by-row Python; GIL contention; slower than equivalent Pandas on single core	Use vectorized Dask operations (`map_partitions`, `assign`, arithmetic operators)
Default blocksize on large CSV files	128MB default creates thousands of partitions for 100GB files; scheduler overhead dominates	Set `blocksize="256MB"` or `blocksize="1GB"` for large files; profile optimal size
`len(df)` without `compute()`	Triggers full dataset read and count; defeats lazy evaluation	Use `df.shape[0].compute()` explicitly; only compute when size is truly needed
Threaded scheduler for CPU-bound work	Python GIL serializes CPU computation across threads; no true parallelism	Use `dask.distributed.LocalCluster()` or process-based scheduler for CPU tasks

Before starting:

cat .claude/context/memory/learnings.md

After completing: Record any new patterns or exceptions discovered.

ASSUME INTERRUPTION: Your context may reset. If it's not in memory, it didn't happen.