Workaround: Workaround Exists
Workaround Description: When using JBDS and using jdbc import to create the a ModeShape model, the duplicate columns should be removed from the following:

Duplicate Column jcr:path in table nt:address
Duplicate Column jcr:name in table nt:childNodeDefinition
Duplicate Column jcr:name in table nt:propertyDefinition
project_key: SOA

When using Designer to import from ModeShape, the table nt_childNodeDefinition, is exposing 2 columns (jcr_name and jcr_name1) pointing to the same underlying sourced column of jcr:name.   This causes an issue when using "Select *" cause is causes miss-alignment of the columns when processing the underlying ModeShape datasource.   And because the columns are not aligned, they don't match the modeshape underlying array's, and therefore, it gets into a situation where the Teiid translator is expecting one data type (i.e., long), but the ModeShape array contains a different one (i.e., String) and it throws this exception: 

Caused by: javax.jcr.ValueFormatException: Cannot convert String value to long
	at org.modeshape.jcr.JcrValue.createValueFormatException(JcrValue.java:100)
	at org.modeshape.jcr.JcrValue.getLong(JcrValue.java:200)
	at org.modeshape.jdbc.JcrResultSet.getValueObject(JcrResultSet.java:1282)
	... 19 more

The short term fix is to change the ModeShape.xmi model in the ModeShape.vdb to remove the jcr_name1 reference in the table.    But need to find out why ModeShape is exposing both columns for the nt_childNodeDefinition table.