All Classes and Interfaces
Class
Description
Standard aggregation operations supported across CupOfTea DataFrames, GroupBy, and pivot operations.
Supported Arrow IPC serialization formats.
Bidirectional converter between CupOfTea
Schema and Apache Arrow Schema.Handles fast, zero-copy buffer mapping between Apache Arrow
FieldVector / ArrowRecordBatch
/ VectorSchemaRoot instances and CupOfTea ColumnVector / IDataFrame instances,
including Apache Arrow C Data Interface export and import.Direction for As-Of Join lookups when matching ordered/temporal keys.
Main command-line launcher for the CupOfTea JMH benchmark suite.
High-performance columnar vector for booleans.
Logical type marker and value representation for dictionary-encoded / categorical string columns.
Alias subclass of
DictionaryColumnVector for Pandas-style Categorical column nomenclature.Converter for
Categorical columns.Declarative chart interface produced by
Plot.Micro-benchmarks comparing sequential row-at-a-time scan speeds vs. vectorized contiguous bulk
operations across primitive ColumnVector implementations.
Fluent reference to a DataFrame column providing ergonomic vectorized string (.str)
and temporal (.dt) operations.
Base abstraction for columnar storage vectors in CupOfTea.
A zero-copy contiguous view over a range
[offset, offset + length) of a ColumnVector.Enterprise compact running GroupBy engine for streaming / out-of-core aggregations.
Comparative benchmarks comparing CupOfTea (sequential and parallel) against Tablesaw
for filtering, aggregation, group-by, and sorting operations.
Axis along which DataFrames are concatenated.
FormatScanner provider for streaming / lazy CSV scans.High-performance streaming CSV data source.
Memory-mapped streaming partition reader for high-performance out-of-core CSV evaluation.
Fallback streaming partition reader using OpenCSV for complex quoting or non-standard encodings.
Marker class for the
cupoftea-all aggregator bundle.Enterprise columnar DataFrame implementation.
High-performance columnar concatenation engine for CupOfTea DataFrames.
Vectorized Columnar Hash Join engine operating directly on
ColumnVector storage.Fluent API for reading data frames from various formats.
Fluent API builder for creating lazy streaming DataFrames from various tabular formats.
Fluent API for writing data frames into various formats.
Dataset scale benchmarks measuring execution speed and scaling across dataset scales
(10k, 100k, 1M, and 10M rows) for single-threaded vs. df.parallel() operations.
Standard logical data types supported by CupOfTea schema and type inference.
Deprecated.
Dictionary / Categorical encoded columnar vector for repeated string values.
High-performance primitive columnar vector for 64-bit IEEE 754 floating point doubles.
Macro-benchmarks for format ingest (CSV, JSON, Parquet, Excel, XML) and
format egress/serialization (CSV, JSON, Parquet).
End-to-end analytical pipeline macro-benchmarks:
Missing value imputation (fillna, dropna) -> Row predicate filter -> Computed column derivation (addColumn) ->
Group-by aggregation (low & high cardinality) -> Parquet / JSON serialization.
Expanding window builder and computation engine.
High-performance un-synchronized direct byte accumulator for parallel CSV serialization.
High-performance byte-level parsing utilities for fast CSV ingestion.
High-performance chunk-based parallel CSV reader using two-stage SIMD structural scanning
(
SimdCsvScanner), direct byte parsing (FastByteParser), zero/low object allocation,
and NIO memory-mapping / buffered block reading, with automatic scalar fallback.Contiguous UTF-8 byte buffer and row offset index for a single chunk of a String column.
A fast, adaptive string cache for CSV decoding to avoid duplicate String instances
for repeated low-cardinality categorical columns (e.g. city names, stations, status codes),
while avoiding allocation overhead on high-cardinality text columns.
High-performance primitive columnar vector for 32-bit IEEE 754 floating point floats.
Service Provider Interface (SPI) for reading tabular data formats.
Registry and runtime discovery hub for format readers, writers, and scanners using
ServiceLoader.Service Provider Interface (SPI) for scanning and streaming tabular data formats lazily.
Service Provider Interface (SPI) for writing tabular data formats.
Deprecated.
Group objects in cells are deprecated and replaced by hash aggregation in GroupBy and Aggregations.
Accumulator for computing group aggregations in a single pass.
Factory and concrete implementations of single-pass group accumulators.
Enterprise columnar GroupBy engine.
High-performance primitive-backed open-addressing hash table for GroupBy partitioning.
Holds group partition state and supports on-demand lazy group indexing and row-map materialization.
Cross-platform JMH
InternalProfiler that captures hardware CPU counters
(Instructions Retired, Cycles Elapsed, IPC) and Peak Resident Set Size (RSS)
per benchmark iteration.Importer for HTML tables into a
CupOfTea data frame.Pure integer-location based indexer for selection by position (similar to pandas iloc).
Represents the index (row labels or designated index column) of a DataFrame.
LazySource backed by an in-memory CupOfTea DataFrame.High-performance primitive columnar vector for 32-bit integers.
Contiguous cache-line interleaved numeric accumulator for grouped aggregations.
High-performance Introselect (hybrid Quickselect with fallback) for linear O(n)
rank selection, median, and quantile calculations on primitive double arrays.
Types of join operations supported between DataFrames.
Configuration options for JSON export.
Jupyter notebook integration helper for Spencer Park's IJava kernel
and other JVM-based notebook environments.
Retention strategy for deduplication operations (e.g., dropDuplicates and distinct).
Fluent Lazy DataFrame evaluation node.
Enterprise lazy / streaming DataFrame evaluation interface for out-of-core data processing.
Lazy GroupBy operation on a
LazyCupOfTea execution plan.Tabular data source capable of streaming partitions/chunks lazily.
Label/Index-based indexer for selection (similar to pandas loc).
High-performance primitive columnar vector for 64-bit integers.
Micro-benchmarks measuring latency and throughput for primitive map-reduce aggregations:
sum, mean, min, max, median, std, and count.
Scaling benchmarks measuring multi-core CPU efficiency across 1, 2, 4, 8, and 16 worker threads
using df.parallel(executor).
Defines the mutation strategy for
CupOfTea transformations.First-class missing-value sentinel, representing an absent / missing value (similar to pandas NA).
High-performance NIO
InputFile backed by a local file and FileChannel.Columnar vector for general reference objects, dates, timestamps, big decimals, and groups.
Standard One Billion Row Challenge (1BRC) benchmark suite for CupOfTea.
Micro-benchmarks comparing single-threaded vs. parallel prefix-sum consolidation
filter performance across varying selectivity rates (1%, 10%, 50%, 90%).
Parallel execution engine for partitioning contiguous
ColumnVector segments across worker threads,
performing parallel prefix-sum (scan) compaction, and executing map-reduce aggregations.Functional interface for a task operating on a chunk
(chunkIndex, startInclusive, endExclusive).High-performance Parquet exporter that consumes primitive ColumnVectors directly
without Avro or Hadoop GenericRecord row-by-row object allocations,
and parallelizes row group encoding across available cores.
FormatScanner provider for lazy Parquet scans with projection pushdown.High-performance streaming and lazy Parquet data source with column projection pushdown,
zero-copy off-heap byte buffer decompression, and lazy predicate pruning.
Sequential reader for a single stream partition or chunk slice.
Declarative plot builder API for DataFrame visualization.
Service Provider Interface (SPI) for pluggable visualization engines.
Registry and runtime discovery hub for plotting providers using
ServiceLoader.Convenient static factory methods for creating plots and charts.
Comparative Arrow RecordBatch & C Data conversion benchmark (CupOfTea vs.
Comparative File Creation & Export (Egress) benchmarks (CupOfTea vs.
Comparative Filter & Vectorized Expression benchmarks (CupOfTea vs.
Comparative GroupBy & Aggregation benchmarks (CupOfTea vs.
Comparative ingestion & parsing benchmarks (CupOfTea vs.
Comparative Join benchmarks (CupOfTea vs.
Comparative Window Functions and Rolling Analytics benchmarks (CupOfTea vs.
Options passed to
FormatReader when ingesting tabular data.Sliding rolling window builder and computation engine.
An immutable representation of a single DataFrame row with typed accessor methods.
Context configuration passed to
LazySource when requesting partitions.Options passed to
FormatScanner when configuring lazy tabular scans.Represents the schema (ordered column names and types) of a DataFrame.
Two-stage SIMD-accelerated CSV structural scanner using the Java Vector API
(
jdk.incubator.vector.ByteVector).Unboxed callback invoked for each active (outside-quote) structural delimiter or newline.
JMH benchmark suite comparing Scalar vs SIMD (
jdk.incubator.vector.ByteVector) CSV scanning and parsing across:
Standard wide CSV: 20 columns, mostly numeric/unquoted values.
High-density quoted CSV: Text fields with embedded commas, escaped quotes, and multiline records.
Mixed CSV with varied row lengths: Row sizes ranging from ~50 bytes to ~2 KB.
Runtime hardware and JVM capability detector for the Java Vector API (
jdk.incubator.vector).Fused Scan-Aggregate execution node for
LazyCupOfTea.Vectorized string expression helper providing fluent text transformations.
High-performance columnar vector for UTF-8 strings backed by an Arrow-compatible
contiguous byte buffer (
byte[] values), an offset index array (int[] offsets),
and a BitSet nulls bitmap.Deterministic synthetic data generator for micro-, macro-, 1BRC, and Polars comparative benchmarks.
Configuration options for rendering DataFrames into rich, responsive HTML tables.
Renders DataFrames into clean, responsive HTML tables with sticky headers,
data type badges, scrolling, and interactive pagination.
Vectorized temporal expression helper providing fluent date and time operations.
Standard TPC-H analytical queries (Q1 - Q22) benchmark suite for CupOfTea.
Deterministic standard TPC-H synthetic dataset generator for benchmarks.
Complete in-memory TPC-H dataset container.
Vega-Lite implementation of
Chart.Concrete implementation of
Plot producing interactive VegaLiteChart instances.Service provider implementation producing Vega-Lite plotting engines.
SQL-like analytical window specification and execution builder.
Direction of observation inclusion for sliding rolling windows.
Options passed to
FormatWriter when exporting tabular data.Micro-benchmarks validating zero-copy slicing and partitioning overhead on ColumnVectors
(sliceView vs slice, partition chunking, and sub-view traversal).
LocalDateConverter,LocalDateTimeConverter, orInstantConverterbased onjava.timeinstead of legacyDate.