Cannot compile parquet-tools

Question

I cloned the repository of parquet-mr from this link. The I wanted to build parquet-tools as explained here:

cd parquet-mr/parquet-tools/
mvn clean package -Plocal

(I updated pom.xml to point correct version of com.twitter:parquet-hadoop:jar)

This is the compilation error that I get:

[ERROR] COMPILATION ERROR :
[INFO] -------------------------------------------------------------
[ERROR] /usr/local/parquet-mr-master/parquet-tools/src/main/java/org/apache/parquet/tools/command/MergeCommand.java:[76,13]
cannot find symbol symbol: method appendFile(org.apache.hadoop.conf.Configuration,org.apache.hadoop.fs.Path) location: variable writer of type org.apache.parquet.hadoop.ParquetFileWriter
[ERROR] /usr/local/parquet-mr-master/parquet-tools/src/main/java/org/apache/parquet/tools/command/MergeCommand.java:[82,29] cannot find symbol symbol: method mergeMetadataFiles(java.util.List,org.apache.hadoop.conf.Configuration) location: class org.apache.parquet.hadoop.ParquetFileWriter

Answer 1

Try checking out a concrete tag in git first.

git checkout apache-parquet-1.9.0
cd parquet-mr/parquet-tools/
mvn clean package -Plocal

By doing the git checkout, all the pom.xml files will say <version>1.9.0</version> instead of <version>1.9.0-SNAPSHOT</version>.