Question

使用Ambari 2.4.0.1的HDP-2.5.0.0

需要将几个SQL Server和Oracle数据库架构导入HDFS / Hive。

目前的方法运作良好：

Sqoop以avro格式从RDBMS导入HDFS
在avro文件顶部创建一个Hive外部表。 dataaggregate_avro_compressed
从步骤2创建最终表格。 自动执行此步骤
将步骤2中的表格中的数据插入最终表格

现在，Step3。 table必须是ORC + COMPRESSED + PARTITIONED ，可能还有MANAGED。手动，可以完成以下操作：

CREATE TABLE `dataaggregate_orc_empty`( ......)PARTITIONED BY (`datedimensionid` bigint)  ROW FORMAT SERDE 'org.apache.hadoop.hive.ql.io.orc.OrcSerde' STORED AS INPUTFORMAT 'org.apache.hadoop.hive.ql.io.orc.OrcInputFormat' OUTPUTFORMAT 'org.apache.hadoop.hive.ql.io.orc.OrcOutputFormat' TBLPROPERTIES ('orc.compress'='ZLIB');

但 AUTOMATIC创建是一项挑战，我正在努力采用以下方法：

CTAS空表

CREATE TABLE dataaggregate_orc_empty LIKE dataaggregate_avro_compressed ROW FORMAT SERDE'org.apache.hadoop.hive.ql.io.orc.OrcSerde'stORED as INPUTFORMAT'org.apache.hadoop.hive.ql.io.orc.OrcInputFormat'OUTPUTFORMAT' org.apache.hadoop.hive.ql.io.orc.OrcOutputFormat'TBLPROPERTIES（'orc.compress'='ZLIB'）;

现在，此表包含分区列datedimensionid ，需要从表中删除，但不支持“alter table drop column”

CTAS使用REGEX column spec。：

set hive.support.quoted.identifiers = none; CREATE TABLE dataaggregate_orc_empty AS SELECT (datedimensionid)?+.+ FROM dataaggregate_avro_compressed limit 0;

这会创建没有分区列datedimensionid的表，但现在如何更改此空表以包含分区列，这就是第一种方法即将到来的地方！ documentation talks about adding partitions有一个规格，但我现阶段没有 - 我只想让这个表与手动创建的表相似（显示在帖子的开头）。 / p>

我该怎么办？

Answer 1

这是连接HiveMetaStoreClient的一种方法，您可以使用方法alter partition。

在此类和列中，可以提取所有其他信息（如分区）。请。请参阅示例客户端和方法。

import org.apache.hadoop.hive.conf.HiveConf;

// test program
public class Test {
    public static void main(String[] args){

        HiveConf hiveConf = new HiveConf();
        hiveConf.setIntVar(HiveConf.ConfVars.METASTORETHRIFTCONNECTIONRETRIES, 3);
        hiveConf.setVar(HiveConf.ConfVars.METASTOREURIS, "thrift://host:port");

        HiveMetaStoreConnector hiveMetaStoreConnector = new HiveMetaStoreConnector(hiveConf);
        if(hiveMetaStoreConnector != null){
            System.out.print(hiveMetaStoreConnector.getAllPartitionInfo("tablename"));
        }
    }
}


// define a class like this

import com.google.common.base.Joiner;
import com.google.common.collect.Lists;
import org.apache.hadoop.hive.conf.HiveConf;
import org.apache.hadoop.hive.metastore.HiveMetaStoreClient;
import org.apache.hadoop.hive.metastore.api.FieldSchema;
import org.apache.hadoop.hive.metastore.api.MetaException;
import org.apache.hadoop.hive.metastore.api.Partition;
import org.apache.hadoop.hive.metastore.api.hive_metastoreConstants;
import org.apache.hadoop.hive.ql.metadata.Hive;
import org.apache.thrift.TException;
import org.joda.time.DateTime;
import org.joda.time.format.DateTimeFormatter;

import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;

public class HiveMetaStoreConnector {
    private HiveConf hiveConf;
    HiveMetaStoreClient hiveMetaStoreClient;

    public HiveMetaStoreConnector(String msAddr, String msPort){
        try {
            hiveConf = new HiveConf();
            hiveConf.setVar(HiveConf.ConfVars.METASTOREURIS, msAddr+":"+ msPort);
            hiveMetaStoreClient = new HiveMetaStoreClient(hiveConf);
        } catch (MetaException e) {
            e.printStackTrace();
            System.err.println("Constructor error");
            System.err.println(e.toString());
            System.exit(-100);
        }
    }

    public HiveMetaStoreConnector(HiveConf hiveConf){
        try {
            this.hiveConf = hiveConf;
            hiveMetaStoreClient = new HiveMetaStoreClient(hiveConf);
        } catch (MetaException e) {
            e.printStackTrace();
            System.err.println("Constructor error");
            System.err.println(e.toString());
            System.exit(-100);
        }
    }

    public String getAllPartitionInfo(String dbName){
        List<String> res = Lists.newArrayList();
        try {
            List<String> tableList = hiveMetaStoreClient.getAllTables(dbName);
            for(String tableName:tableList){
                res.addAll(getTablePartitionInformation(dbName,tableName));
            }
        } catch (MetaException e) {
            e.printStackTrace();
            System.out.println("getAllTableStatistic error");
            System.out.println(e.toString());
            System.exit(-100);
        }

        return Joiner.on("\n").join(res);
    }

    public List<String> getTablePartitionInformation(String dbName, String tableName){
        List<String> partitionsInfo = Lists.newArrayList();
        try {
            List<String> partitionNames = hiveMetaStoreClient.listPartitionNames(dbName,tableName, (short) 10000);
            List<Partition> partitions = hiveMetaStoreClient.listPartitions(dbName,tableName, (short) 10000);
            for(Partition partition:partitions){
                StringBuffer sb = new StringBuffer();
                sb.append(tableName);
                sb.append("\t");
                List<String> partitionValues = partition.getValues();
                if(partitionValues.size()<4){
                    int size = partitionValues.size();
                    for(int j=0; j<4-size;j++){
                        partitionValues.add("null");
                    }
                }
                sb.append(Joiner.on("\t").join(partitionValues));
                sb.append("\t");
                DateTime createDate = new DateTime((long)partition.getCreateTime()*1000);
                sb.append(createDate.toString("yyyy-MM-dd HH:mm:ss"));
                partitionsInfo.add(sb.toString());
            }

        } catch (TException e) {
            e.printStackTrace();
            return Arrays.asList(new String[]{"error for request on" + tableName});
        }

        return partitionsInfo;
    }

    public String getAllTableStatistic(String dbName){
        List<String> res = Lists.newArrayList();
        try {
            List<String> tableList = hiveMetaStoreClient.getAllTables(dbName);
            for(String tableName:tableList){
                res.addAll(getTableColumnsInformation(dbName,tableName));
            }
        } catch (MetaException e) {
            e.printStackTrace();
            System.out.println("getAllTableStatistic error");
            System.out.println(e.toString());
            System.exit(-100);
        }

        return Joiner.on("\n").join(res);
    }

    public List<String> getTableColumnsInformation(String dbName, String tableName){
        try {
            List<FieldSchema> fields = hiveMetaStoreClient.getFields(dbName, tableName);
            List<String> infs = Lists.newArrayList();
            int cnt = 0;
            for(FieldSchema fs : fields){
                StringBuffer sb = new StringBuffer();
                sb.append(tableName);
                sb.append("\t");
                sb.append(cnt);
                sb.append("\t");
                cnt++;
                sb.append(fs.getName());
                sb.append("\t");
                sb.append(fs.getType());
                sb.append("\t");
                sb.append(fs.getComment());
                infs.add(sb.toString());
            }

            return infs;

        } catch (TException e) {
            e.printStackTrace();
            System.out.println("getTableColumnsInformation error");
            System.out.println(e.toString());
            System.exit(-100);
            return null;
        }
    }
}

更改（EMPTY）表添加分区COLUMN

1 个答案: