From scratch Heap file and Hash index #

Originally written for some undergraduate DB systems class I took.

This is still a bit interesting to me as it’s a from-scratch implementation of part of a DB and I think it would be a good exercise to revisit this code and fix the bugs that are definitely in there somewhere.

Will be funny to see how bad the Java code is too.

Usage #

The dataset used in this project is available at https://data.gov.au/data/dataset/asic-business-names but I’ve also copied a subset of that dataset into ./data/

dbload #

Implementing a heap file in Java. Load a database relation (.tsv) and write a heap file.

bazel run //databases/fromscratch_heapfile_and_hashindex:dbload -- \
    -p 4096 \
    "$(git rev-parse --show-toplevel)/databases/fromscratch_heapfile_and_hashindex/data/TRUNCATED_DATASET_WO_HEADER.csv"

hashload #

A hash indexer that uses the heap file to build an index, hash.<pagesize>.

dbquery #

Perform a text search using the heap file, with or without an index.

Build #

bazel build //databases/fromscratch_heapfile_and_hashindex/...

Tests #

… I guess tests weren’t part of the assignment.