Overview

Jawk is a pure Java implementation of AWK^[1]. You can run AWK programs in a JVM application and integrate with your Java code:

$ echo "hello world" | jawk '{ print $2 ", " $1 "!" }'
world, hello!

Awk awk = new Awk();
String result = awk.script("{ print toupper($0) }").input("hello world").execute();
// result = "HELLO WORLD\n"

Compatibility and Compliance

The above scores are calculated from the latest compatibility tests during mvn verify site. See the full compatibility dashboard^[2] for per-suite maps and the Failsafe report^[3] for detailed test results.

Key Capabilities

• A drop-in AWK command line: inline scripts, script files, name=value operands, and input files
• Java entry points to run a whole script, evaluate a single expression, or compile either one once and reuse it
• Structured input through InputSource, to feed rows or already-tokenized fields straight from your own data structures
• Custom output through AwkSink, to collect what a script prints as Java values instead of text
• Gawk builtins enabled by default: asort(), asorti(), typeof(), isarray(), mkbool(), gensub(), patsplit(), strtonum(), systime(), mktime(), strftime(), the gettext functions, and PROCINFO["sorted_in"]
• Java extensions, loaded from the CLI or the Java API, to expose your own functions to a script
• A sandboxed mode, in Java and on the CLI, that rejects system(), redirections, pipes, and @include
• Precompilation: -K writes a compiled program to a file, -L runs it later without parsing the source again

Differences with Traditional AWK

Jawk aims to be a practical AWK implementation for JVM environments, but it is not a byte-for-byte clone of every historical AWK behavior. Some differences are deliberate and come from the way Jawk integrates with Java:

• Regular expression behavior follows Java's regex engine, which may differ from traditional AWK regexes in edge cases. Notably, alternation picks the first matching branch rather than the POSIX longest one, which can affect match(), field splitting with patsplit(), and similar content-driven matching: order alternatives longest-first.
• printf() and sprintf() implement the POSIX AWK conversions with gawk's semantics, including CONVFMT-based %s conversion, dynamic * width and precision, and gawk's out-of-range integer handling; only a few edge cases (%a notation, %c locale independence, the sign of NaN) follow Java's platform rules.
• Some floating-point edge cases may differ due to Java's handling of floating-point arithmetic and representation.
• Where AWK ignores whitespace around numbers — the leading whitespace in string-to-number conversion, and the leading and trailing blanks that POSIX lets a numeric string carry — Jawk applies Java's whitespace definition (Character.isWhitespace), which also covers Unicode spaces such as U+2003; C-based implementations only skip the C locale's narrower isspace set.
• The date and time functions (mktime(), strftime()) follow the Java platform's time zone data and calendar rules rather than the C library's, which differs from gawk in a few edge cases.

The Compatibility^[2] page details these edge cases, along with the places where Jawk deliberately follows gawk rather than historical AWK, and the live compatibility reports.

History of Jawk

Next Steps

• Install Jawk^[17]
• Learn the CLI^[18]
• Embed Jawk in Java^[19]
• Load or write extensions^[20]