Source: Oracle table with some columns defined as Number without precision and scale. But the actual value could contain both integer or decimal values e.g 4.51 or 4.

Target: parquet file

ETL tool: informatica developer

Informatica pdo is the oracle table (relational object), and the column definition becomes DOUBLE for those Number columns. in the INFA mapping expression, I need to convert these double to decimal in order to preserve the decimal values as mentioned above.

But when I create the hive table on top of this parquet file, I get the error ‘org.apache.parquet.io.ParquetDecodingException: Can not read value at 0 in block -1 in file’

I know there’s a solution which I need to set the –conf “spark.sql.parquet.writeLegacyFormat=true”. But this is a spark configuration which I can’t do, because I’m using ‘Native’ mode in the infa mapping. (spark mode doesn’t support oracle relational object)

So what can I do?