Getting Started¶
- What is it?
- Python-Blosc2’s compute engine
- Data Containers
- Computing with NDArrays
- Querying Columnar Data with CTable
- Installation
- Tutorials
- NDArray: A NDim, Compressed Data Container
- Expressions containing NDArray objects
- User Defined Functions
- LazyArray UDF DSL Kernels
- Optimizing data reductions with NDArrays
- Advanced Lazy Expressions and Persistent Reductions
- Using Proxies for Efficient Handling of Remote Multidimensional Data
- Compressing data with the SChunk class
- Slicing, extending and serializing with SChunks
- User-defined codecs and filters
- Prefilters, postfilters and fillers
- Working with Containers
- Working with ObjectArray
- Working with BatchArray
- CTable Tutorial
- Indexing Arrays
- Indexing CTables
- miniexpr DSL Syntax (Canonical Reference)
- Quick start
- Program shape
- Header pragmas
- Function signature and inputs
- Statements
- Expressions and function calls
- Temporary variable type inference
- Loops
print(...)- Reserved names
- ND reserved symbols
- Typing and return behavior
- Compound assignment desugaring
- Compile-time vs runtime errors
- Python syntax that is out of DSL scope
- Parquet to Blosc2 Walkthrough
- Prerequisites
- Step 1 — Create a sample Parquet file
- Step 2 — Import to a compact
.b2zstore - Step 3 — Import to a sparse
.b2dstore - Step 4 — Fixed-width string import
- Step 5 — Custom chunk and block layout
- Step 6 — Disable the summary index
- Step 7 — Export back to Parquet
- Step 8 — Spot-check the exported file
- b2view: Browse TreeStore Bundles in the Terminal