Enhanced searches with fzf

Command Line – fzf

© Photo by Timothy Eberly on Unsplash

© Photo by Timothy Eberly on Unsplash

Article from Issue 268/2023

Simplify your searches and get better results with fzf, a modern search tool based on fuzzy logic.

Search commands have always been essential to using computers. With the increased storage capacity of modern systems, they are more important than ever because there is more material to search. In fact, searches are so important that alternatives to time-honored commands are becoming increasingly common. One of the most effective of these recent alternatives is fzf, a command-line "fuzzy finder" [1].

For most Linux users, the basic search tool for finding directories and files has been find, which is most effective when you know exactly what you are looking for. If you do not know, then you need a way to allow variation in the output. Traditionally, variation in search tools like find is provided by regular expressions (regexes), a concept first developed in the 1950s. Regexes are standard patterns that offer different types of variations in results. Most users are familiar with the simpler regexes, such as * to mean zero or any character, so that apt* might return apt, apt-get, apt-cache, and probably many others. At a more advanced level, users might know that gray|grey produces results with either spelling, or [b-e] a range of characters. However, regexes are not consistent across applications and can rapidly become extremely complex to construct or type correctly, especially when all you want is a quick answer. For example, even with context, who can easily explain this randomly selected example:


Personally, I'd take half a minute or so and resent the expected effort. Such regexes are worth constructing only when you use the same search repeatedly. In addition, you almost always will find scanning fuzzy logic results quicker than entering a lengthy regex, even if you type the regex correctly the first time. With fzf, you interact with the tool through a fuzzy logic interface, while fzf composes the regexes for a search behind the scenes.

Fuzzy logic is a science of its own, with principles and theories about communication and logic that do not directly concern us here [2]. What matters is that fuzzy logic is a much simpler way of introducing variation into output, especially today when the patience and general command-line knowledge among computer users is much less than it was 70 years ago. For one thing, the order in which variation is found with fuzzy logic is less important than with regexes. Often, the order does not matter at all, which makes it easier to understand and more flexible. Just as importantly, fuzzy logic uses natural language instead of its own code that needs to be carefully separated from the query itself. If you add an indexed database, as fzf does, you can add context to accuracy to produce a contextual command history.

With these advantages, fzf has rapidly gained ground in the past few years. It is carried in all the major distributions, or it can be installed from its GitHub page. Each time fzf opens using the bare command, it updates its database in a matter of seconds, leaving you at a prompt, ready to continue.

Setting Up fzf

The popularity of fzf means that, for most purposes, you can install it directly from your distribution's repositories and begin to use it. You can also install fzf with Homebrew to get the absolute latest version.

However, to make full use of fzf, you should also install bat (the modern cat replacement) and fd-find (the modern find replacement). Then add the following lines to your .bash.rc files:

export FZF_DEFAULT_OPTS="--preview 'bat --color=always {}'"
export FZF_DEFAULT_COMMAND="fd --type f"

These lines will ensure that fzf can use bat as a file previewer and improve overall performance with fd-find. Faster than find, fd-find will ignore directories like .gitignore while searching. In addition, add

$ source /usr/share/fzf/shell/key-bindings.bash

to enable some useful keyboard shortcuts (see Table 1).

Table 1

Keyboard Shortcuts


Select files


Switch into a selected subdirectory


Use fzf's enhanced history

Basic Use

At its simplest, fzf opens and updates its list of top-level directories for the present working directory – usually, your home directory. You can then type any part of a directory or file to see the results (Figure 1). If you scroll the results, highlighting and pressing the Enter key displays the selection in the terminal, where it can be conveniently pasted into newly typed commands. Alternatively, you can use fzf with another command so that

vim $(fzf)

allows you to use fzf to select a file to open in Vim.

Figure 1: The results shown in the fzf window.

For many uses, this is enough information to use fzf. However, depending on how fzf is compiled, it can also support context-specific autocompletions by entering a command followed by **+Tab and will give appropriate suggestions, such as file names for Vim or remote systems for SSH. If autocompletion is not enabled, check the project's GitHub page for instructions.

Other Command Options

Most online articles on fzf end at this point. However, a few distributions, such as Ubuntu, offer man pages with additional options, neatly organized into categories. Among the most immediately useful are the search modes. Some of these options may be familiar to you from find and other search commands. For instance, with +1, searches become case-sensitive, while you can reduce clutter in the results with --exact (-e). In addition, --algo=TYPE sets the fuzzy algorithm, completion with v1 optimizes speed over best results, and v2 optimizes results over speed.

Another useful option category is for search results. The arrangement of results can be influenced by --tiebreak=CRI CRITERIA, CRITERIA. The default criteria is length, which favors shorter results, but you can give it lower priority by its position in a list that can contain begin (with the search string closer to the start of the line), end (with the search string closer to the end of the line), or index (a line that already appears in the results). Each criterion should only appear once, and index (for obvious reasons) must always appear last. If you are searching for high-level directories, start might be a useful option or end if you are searching for a file name. In addition, --tac displays results in reverse order.

Other options are cosmetic, such as disabling the mouse or horizontal or vertical control, or aspects of appearance, such as color theme or window border size for previews. These options no doubt become more important the more often you use fzf, but at first glance give a false sense of complexity for what is after all a relatively well-organized utility.

Buy this article as PDF

Express-Checkout as PDF
Price $2.95
(incl. VAT)

Buy Linux Magazine

Get it on Google Play

US / Canada

Get it on Google Play

UK / Australia

Related content

  • Fzf/Fzy

    Fuzzy finders retrieve useful results from data streams even if there are no exact matches.

  • Tracked Down

    Searching for text in files or data streams is a common and important function. Ugrep tackles this task quickly, efficiently, and even interactively if needed.

  • Regex Generators

    As regular expressions grow in complexity, regex generators can make the job easier by computing the patterns for you.

  • agrep

    The agrep tool expands on grep by adding fuzzy search capabilities to text string-matching operations.

  • Command Line – tre-agrep

    Tre-agrep has all of grep's functionality but can also do ambiguous or fuzzy searches without deep knowledge of regular expressions.

comments powered by Disqus
Subscribe to our Linux Newsletters
Find Linux and Open Source Jobs
Subscribe to our ADMIN Newsletters

Support Our Work

Linux Magazine content is made possible with support from readers like you. Please consider contributing when you’ve found an article to be beneficial.

Learn More