Data engineering · science · architecture

IngestThis

Working notes for people who move data for a living. Pipelines, lakehouses, table formats, and the architecture underneath them.

Read the archiveJoin the Slack
Articles
450+
Topics
Iceberg, pipelines, AI
Submissions
Open
Price
Free

Latest

All articles
  1. 01Apache Arrow Flight and ADBC, and Why Database Connectivity Finally Went Columnar2026-08-06
  2. 02Budgeting for Agentic Analytics When Every Question Costs Something DifferentBudgeting for agentic analytics when every question costs something different: token economics, query economics, instrumentation, and the cost controls that actually return.2026-08-04
  3. 03The Five Layers of an Agentic Lakehouse and Where the MCP Server SitsThe five layers of an agentic lakehouse and where the MCP server sits: storage, catalog, semantic layer, MCP gateway, and agent surface, plus identity, session isolation, and budgets.2026-08-04
  4. 04Autonomous Table Optimization When Your Query Workload Stops Being PredictableAutonomous table optimization when query workloads stop being predictable: observing file layout and query patterns, scoring compaction work, adaptive sort order, and cost discipline.2026-08-04
  5. 05Building Apache Iceberg Lakehouses That Run Without an Internet ConnectionHow to build an Apache Iceberg lakehouse that runs fully offline: storage, catalog, compute, cross-zone transfer, compliance, and the failure modes that bite.2026-08-04
  6. 06Wiring Analytical Queries to Transactional APIs in Closed-Loop Decision AgentsWiring analytical queries to transactional APIs in closed-loop decision agents: conditional writes, sagas with compensations, decision records, and blast radius controls.2026-08-04
  7. 07Surviving Optimistic Commit Collisions When Hundreds of Agents Write to IcebergSurviving optimistic commit collisions when hundreds of agents write to Iceberg: which conflicts are real, commit buffers, partitioning, and the patterns that prevent commit storms.2026-08-04
  8. 08How Iceberg V3 Deletion Vectors Fixed Merge-on-Read for Streaming TablesHow Iceberg V3 deletion vectors replaced accumulating positional delete files and made merge-on-read viable for streaming and CDC tables.2026-08-04

Reference shelf

Long reads from around the web
Semantic Layer

The Semantic Layer: Definitive Guide

A comprehensive guide to the Semantic Layer — how it creates a single source of truth for metrics, powers headless BI, and makes AI agents answer business questions accurately.

ReadApache Polaris

Apache Polaris: The Catalog Standard for Lakehouses and AI

How Apache Polaris is emerging as the universal Iceberg catalog standard, enabling multi-engine interoperability and governed AI access across the lakehouse ecosystem.

ReadTable Formats

What Are Table Formats and Why Were They Needed?

The origin story of open table formats — the problems with Hive, why Apache Iceberg, Delta Lake, and Hudi were created, and what they unlock for modern data platforms.

ReadDremio

What Is Dremio?

A clear-eyed breakdown of what Dremio is, how its semantic layer, query federation, Reflections, and Apache Arrow Flight power the Intelligent Lakehouse Platform.

ReadApache Iceberg

What Apache Iceberg Native Actually Means

Not all 'Iceberg support' is equal. This piece breaks down what it means to be genuinely Apache Iceberg native versus bolt-on, and why it matters for your lakehouse.

ReadOpen Source

Open Source and the Data Lakehouse

How the Apache Software Foundation's open-source projects — Iceberg, Arrow, Parquet, Polaris — form the modular foundation of the modern open data lakehouse.

ReadAgentic AI

What Is Agentic Analytics?

Agentic AI is reshaping how organizations interact with data. This guide explains agentic analytics, the role of the semantic layer, and why query performance matters for AI agents.

ReadData Lakehouse

Definitive Guide to the Data Lakehouse

The complete, authoritative guide to the Data Lakehouse architecture — what it is, why it supersedes the data warehouse + data lake combination, and how to build one.

ReadAI & Performance

How Dremio Keeps Agentic Analytics Fast Without Manual Tuning

How Dremio's layered autonomous performance architecture — Reflections, caching, vectorized execution — handles unpredictable AI agent query patterns at interactive speed.

Read

Elsewhere

Pitch an idea: alex@ingestthis.com