What's New in Cloudera Search

The MapReduceIndexerTool now supports ingesting parquet DECIMAL data into Solr

The ReadAvroParquetFile morphline command got a new attribute, decimalConversionEnabled. When set to true, decimal Parquet data is correctly read (instead of returning raw bytes as earlier). Its default value is false for backwards compatibility.

Decimal data can be stored as a string, float or double Solr field. Only storage as string is lossless, float and double can result in loss of precision.

Optimize phase of MRIT reducers can now be skipped

This release adds a new feature to MapReduceIndexerTool (MRIT) that allows skipping the optimize phase of reducers. To skip the optimize operation at the end of reducers, set the value of the maxSegments parameter to 0 or -1.