plocate

plocate is a drop-in replacement for the traditional locate command that uses modern data structures and I/O techniques to search through millions of files in milliseconds. While the original locate and its successor mlocate perform linear scans through the entire database, plocate uses an inverted index based on trigrams (three-byte sequences) to rapidly narrow down candidates to a tiny subset before returning results. This approach provides 10-100x faster searches while using significantly less disk space for the index.

The performance improvements are dramatic: plocate can complete a file search in mere milliseconds where mlocate takes seconds. Beyond speed, plocate introduces a more intuitive search behavior: when searching for multiple strings, it returns only files matching all the search terms, rather than any file matching any term. The database format is compatible with mlocate, making migration seamless. For maximum performance on modern Linux systems with kernel 5.1+, plocate uses asynchronous I/O through io_uring to minimize seek latency, especially beneficial on older hard drives.

plocate is suitable as a system-wide replacement for locate, and most Linux distributions now ship it as their default locate implementation. Installation automatically builds the initial database in /var/lib/plocate, and the database is automatically updated through the system's scheduled tasks.

Basic usage

# Search for a file by name (fastest search).
plocate filename

# Search for multiple terms (all must match).
plocate file pattern

# Limit results to a specific directory pattern.
plocate /home/user/file

# Use regular expressions (slower, but more flexible).
plocate -r 'pattern.*file'

# Count matching files.
plocate -c filename