"ParseException" Error is Received on ODI 12c Hive Table Creation when Column Name Has Diacritics

Last updated on JANUARY 05, 2018

Applies to:

Symptoms

Having an Oracle Data Integrator (ODI) 12c Mapping to load data from an Oracle table into Hive.

The Oracle source table has column names with diacritics also known as extended ASCII or accented characters, such as:

     REDEN_COMMERCIËLE_KORT
     (Observe the diacritic Ë)

The following error is reported:

• on the LKM step when creating the C$ table on Hive side...
• on the IKM step when creating the target table on Hive side...

ODI-1228: Task Create work table-LKM SQL to Hive SQOOP- fails on the target connection BDC_Hive.
Caused By: java.sql.SQLSyntaxErrorException: [FMWGEN][Hive JDBC Driver][Hive]Error while compiling statement: FAILED: ParseException line 156:16 cannot recognize input near 'LE_KORT' 'STRING' ',' in column type
  at weblogic.jdbc.hivebase.ddcp.b(Unknown Source)
  at weblogic.jdbc.hivebase.ddcp.a(Unknown Source)
  ...

     ... the executed code being:

CREATE TABLE IF NOT EXISTS odi_work_schema.c_0EXPRESSIONAWknl9hGIbdhJPMDjwBQuj
(
  ...
  REDEN_COMMERCIËLE_KORT STRING,
  ...
)

Cause