Summary of problems encountered in sqoop2 1.99.6 and source code modification
1. When the cardinality of PartitionColumn is 1 (as shown in the following figure). Will report an error.
The error is as shown in the following picture
The source code is as follows
\ sqoop-1.99.6-bin-hadoop200\ connector\ connector-generic-jdbc\ src\ main\ java\ org\ apache\ sqoop\ connector\ jdbc\ GenericJdbcPartitioner.java
Change the source code to
Can
The parameter Null value allowed for the partition column: false in 2.Job will not report an error even if partition column has null, so the record for null is imported into HDFS.
Conclusion: it is not a problem. When this parameter is false, the map of pid is null will not be de-derived, but only when it is true, so it is correct not to report an error.
3. The problem of date and time. The specific modification method is as follows.
E:\ IdeaProject\ sqoop-1.99.6-bin-hadoop200\ connector\ connector-generic-jdbc\ src\ main\ java\ org\ apache\ sqoop\ connector\ jdbc\ GenericJdbcExtractor.java
The time type in 4.oracle will report an error when doing pid.
If (JDBC_DRIVER_ORACLE.equals (jdbcDriver)) {
/ / if it is oracle, do special treatment.
Conditions.append ("to_timestamp (\'") .append (sdf.format ((java.util.Date) lowerBound)) .append ("\', 'yyyy-mm-dd hh34:mi:ss.ff')")
Conditions.append ("