----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/14779/#review27358 -----------------------------------------------------------
src/java/org/apache/sqoop/mapreduce/DataDrivenImportJob.java <https://reviews.apache.org/r/14779/#comment53221> In an updated patch, I've modified the code such that it follows the same pattern as ClassWriter/CompilationManager. It will initially write the *.avsc file to the JAR output directory, and then attempt to move the file to the code output directory (CWD). I thought this was a safe approach just in case there comes a point where having an updated *.avsc file alongside an updated *.java/*.jar file in the JAR output directory is important to someone's Sqoop workflow. The *.avsc file in the code output directory will not be overwritten. - James Anderson On Oct. 20, 2013, 8:10 p.m., James Anderson wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/14779/ > ----------------------------------------------------------- > > (Updated Oct. 20, 2013, 8:10 p.m.) > > > Review request for Sqoop. > > > Bugs: SQOOP-435 > https://issues.apache.org/jira/browse/SQOOP-435 > > > Repository: sqoop-trunk > > > Description > ------- > > Automatically generate an Avro schema file (*.avsc) in the JAR output > directory when --as-avrodatafile is specified. > > > Diffs > ----- > > src/java/org/apache/sqoop/mapreduce/DataDrivenImportJob.java 5afd90c > > Diff: https://reviews.apache.org/r/14779/diff/ > > > Testing > ------- > > - Compared contents of *.avsc file against output from avro-tools getschema > operation. > - Created an external Hive table with AvroSerDe, pointing avro.schema.url to > a copy of the schema indexed in ElasticSearch. > > > Thanks, > > James Anderson > >
