读取oralce 含blob 字段的表会报java.sql.SQLException: 流已被关闭

读取oralce 含blob 字段的表会报java.sql.SQLException: 流已被关闭，导致无法读取源表所有的字段，例如源表有5个字段，只会读取2个（读到blob字段后报错，自动放弃其他字段）
详细报错信息： java.sql.SQLException: 流已被关闭
2020-10-10 16:42:51 [AnalysisStatistics.analysisStatisticsLog-53] at oracle.jdbc.driver.LongAccessor.getBytesInternal(LongAccessor.java:156) ~[ojdbc6-11.2.0.3.jar:11.2.0.3.0]
2020-10-10 16:42:51 [AnalysisStatistics.analysisStatisticsLog-53] at oracle.jdbc.driver.LongAccessor.getBytes(LongAccessor.java:126) ~[ojdbc6-11.2.0.3.jar:11.2.0.3.0]
2020-10-10 16:42:51 [AnalysisStatistics.analysisStatisticsLog-53] at oracle.jdbc.driver.LongAccessor.getString(LongAccessor.java:201) ~[ojdbc6-11.2.0.3.jar:11.2.0.3.0]
2020-10-10 16:42:51 [AnalysisStatistics.analysisStatisticsLog-53] at oracle.jdbc.driver.T4CLongAccessor.getString(T4CLongAccessor.java:450) ~[ojdbc6-11.2.0.3.jar:11.2.0.3.0]
2020-10-10 16:42:51 [AnalysisStatistics.analysisStatisticsLog-53] at oracle.jdbc.driver.OracleResultSetImpl.getString(OracleResultSetImpl.java:1297) ~[ojdbc6-11.2.0.3.jar:11.2.0.3.0]
2020-10-10 16:42:51 [AnalysisStatistics.analysisStatisticsLog-53] at com.alibaba.datax.plugin.rdbms.reader.CommonRdbmsReader$Task.buildRecord(CommonRdbmsReader.java:257) [plugin-rdbms-util-0.0.1-SNAPSHOT.jar:na]
2020-10-10 16:42:51 [AnalysisStatistics.analysisStatisticsLog-53] at com.alibaba.datax.plugin.rdbms.reader.CommonRdbmsReader$Task.transportOneRecord(CommonRdbmsReader.java:237) [plugin-rdbms-util-0.0.1-SNAPSHOT.jar:na]
2020-10-10 16:42:51 [AnalysisStatistics.analysisStatisticsLog-53] at com.alibaba.datax.plugin.rdbms.reader.CommonRdbmsReader$Task.startRead(CommonRdbmsReader.java:209) [plugin-rdbms-util-0.0.1-SNAPSHOT.jar:na]
2020-10-10 16:42:51 [AnalysisStatistics.analysisStatisticsLog-53] at com.alibaba.datax.plugin.reader.oraclereader.OracleReader$Task.startRead(OracleReader.java:110) [oraclereader-0.0.1-SNAPSHOT.jar:na]
2020-10-10 16:42:51 [AnalysisStatistics.analysisStatisticsLog-53] at com.alibaba.datax.core.taskgroup.runner.ReaderRunner.run(ReaderRunner.java:57) [datax-core-0.0.1-SNAPSHOT.jar:na]

Use OracleReader$Task.startRead and CommonRdbmsReader$Task.buildRecord at the stack-trace lines as entry points. Reproduce against an Oracle table containing a BLOB and inspect how columns are read after the BLOB. Done means every source column is read without java.sql.SQLException: Stream closed.

Written by the indexing model from the issue text.