class Hive < Formula desc "Hadoop-based data summarization, query, and analysis" homepage "https://hive.apache.org" url "https://www.apache.org/dyn/closer.lua?path=hive/hive-3.1.2/apache-hive-3.1.2-bin.tar.gz" mirror "https://archive.apache.org/dist/hive/hive-3.1.2/apache-hive-3.1.2-bin.tar.gz" sha256 "d75dcf36908b4e7b9b0ec9aec57a46a6628b97b276c233cb2c2f1a3e89b13462" license "Apache-2.0" revision 1 livecheck do url :stable end bottle :unneeded depends_on "hadoop" depends_on java: "1.8" def install rm_f Dir["bin/*.cmd", "bin/ext/*.cmd", "bin/ext/util/*.cmd"] libexec.install %w[bin conf examples hcatalog lib scripts] # Hadoop currently supplies a newer version # and two versions on the classpath causes problems rm libexec/"lib/guava-19.0.jar" guava = (Formula["hadoop"].opt_libexec/"share/hadoop/common/lib").glob("guava-*-jre.jar") ln_s guava.first, libexec/"lib" Pathname.glob("#{libexec}/bin/*") do |file| next if file.directory? (bin/file.basename).write_env_script file, Language::Java.java_home_env("1.7+").merge(HIVE_HOME: libexec) end end def caveats <<~EOS Hadoop must be in your path for hive executable to work. If you want to use HCatalog with Pig, set $HCAT_HOME in your profile: export HCAT_HOME=#{opt_libexec}/hcatalog EOS end test do system bin/"schematool", "-initSchema", "-dbType", "derby" assert_match "Hive #{version}", shell_output("#{bin}/hive --version") end end