chiudaniels commented on code in PR #16961: URL: https://github.com/apache/iceberg/pull/16961#discussion_r3612021901
########## format/index.md: ########## @@ -0,0 +1,390 @@ +--- +title: "Index Spec" +--- +<!-- + - Licensed to the Apache Software Foundation (ASF) under one or more + - contributor license agreements. See the NOTICE file distributed with + - this work for additional information regarding copyright ownership. + - The ASF licenses this file to You under the Apache License, Version 2.0 + - (the "License"); you may not use this file except in compliance with + - the License. You may obtain a copy of the License at + - + - http://www.apache.org/licenses/LICENSE-2.0 + - + - Unless required by applicable law or agreed to in writing, software + - distributed under the License is distributed on an "AS IS" BASIS, + - WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + - See the License for the specific language governing permissions and + - limitations under the License. + --> + +# Iceberg Index Specification + +## Background and Motivation + +Indexes enable query engines to locate relevant rows without scanning entire datasets. +They can accelerate point lookups, range predicates, and other retrieval patterns +while preserving Iceberg's table format, snapshot isolation, and interoperability. + +Indexes are optional. Engines may choose to create, maintain, consume, or ignore them. + +## Goals + +- Define a portable metadata format for indexes +- Provide a common storage architecture for index data +- Allow indexes to be operated independently from source table metadata +- Enable index sharing across engines +- Provide a framework for defining new index types and transform functions + +## Overview + +Indexes are stored as a collection of files with some Iceberg table like semantics. At a high level they consist of a tracking file (similar to a root manifest file) which contains listings for a defined set of leaf files (similar to data files). Leaf files store an ordered set of rows, each containing at least a key, the path of the Iceberg table data file, and the position within that file where the row for that key is stored. The organization of leaf files is defined by an Index Transform Function which varies based on the type of index. This structure is recorded in an Index metadata.json file which contains a set of snapshots, each of which points to a single tracking file mapping to the complete state of an Iceberg table at a given Iceberg table snapshot. + +Like Iceberg tables, views, and functions: + +- Metadata files (index metadata and tracking files) and data files (leaf files) are immutable +- Updates create new metadata files +- Catalogs perform atomic metadata swaps + +Each index snapshot references a tracking file which describes the leaf files belonging to the snapshot. + +```text +Index Metadata + | + +-- Index Snapshot(s) + | + +-- Tracking File + | + +-- Leaf Data Files +``` + +Transform functions derive a transform value from the key columns and determine how index entries are organized within +the leaf files. +- The transform value space is divided into non-overlapping ranges. +- Each leaf file stores entries for a single range. +- The tracking file stores range bounds for each leaf file. + +This structure enables efficient planning while keeping the data layout flexible for different index implementations. Review Comment: Based on the flat structure and immutability of the of index metadata files, it seems this is suited for flat index structures like hash-based indexing. However, for common multi-level indexing like b+ tree do we intend for flexible implementations of the physical index data and mutability? in the case of immutability, lsm trees might better fit, however would still require leveled index organization. are these types of index implementations currently out of scope for this spec? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
