DataFrames & Tabular Subsystem

High-performance Arrow-backed tabular data processing, CSV, JSON, and Parquet ETL.

NextViper Data Subsystem

The NextViper data package provides columnar DataFrame processing powered by Apache Arrow memory layouts with zero-copy vectorization.

1. Loading Datasets

nextviper

import data

// Load CSV
let df = data.read_csv("telemetry.csv")

// Load Parquet
let parquet_df = data.read_parquet("features.parquet")

print("Dataset Dimensions:", df.shape)
print("Columns:", df.columns())
print(df.head(5))

2. Column Filtering & Transformations

nextviper

// Filter rows where score > 0.85 and status == 'active'
let filtered = df.filter(
    (df["score"] > 0.85) and (df["status"] == "active")
)

// Add computed column
let enriched = df.with_column("normalized_loss", df["loss"] / 100.0)

// Sort values
let sorted_df = enriched.sort_by("score", descending: true)

3. GroupBy & Aggregations

nextviper

let summary = df.group_by("category").aggregate({
    "revenue": "sum",
    "score": "mean",
    "user_id": "count"
})

print(summary)