Skip to main content

Hive: Write custom serde

By Dinesh
Suppose we have input file like below:
$ vi uwserde
kiju1233,1234567890
huhuhuhu,1233330987
…
…
…
This input file consist of sessionid and timestamp as comma-separated value. Assuming this I wrote a WritableComparable as below:

package hive;

import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.io.WritableComparable;

public class UserWritable implements WritableComparable<UserWritable>  {
      
       private Text sessionID;
      
      
       private Text timestamp;


       public UserWritable() {
              set(new Text(), new Text());
       }
      
      
       public void set(Text sessionID, Text timestamp) {
              this.sessionID = sessionID;
              this.timestamp = timestamp;
       }

      
       public Text getSessionID() {
              return sessionID;
       }
      
      
       public Text getTimestamp() {
              return timestamp;
       }
      
      
       @Override
       public void readFields(DataInput din) throws IOException {
              sessionID.readFields(din);
              timestamp.readFields(din);
       }

       @Override
       public void write(DataOutput dout) throws IOException {
              sessionID.write(dout);
              timestamp.write(dout);
       }

      
       @Override
       public int hashCode() {
              return sessionID.hashCode()*163 + timestamp.hashCode();
       }
      
       @Override
       public int compareTo(UserWritable uw) {
              int cmp = compare(sessionID,uw.sessionID); //sessionID.compareTo(uw.sessionID);
             
              if(cmp != 0){
                     return cmp;
              }
             
              try{
                     long meT = Long.parseLong(timestamp.toString());
                     long cmpT = Long.parseLong(uw.timestamp.toString());
                     return compare(new LongWritable(meT),new LongWritable(cmpT));//new LongWritable(meT).compareTo(new LongWritable(cmpT));
              }catch(Exception e){
                     e.printStackTrace();
                     throw new RuntimeException("Error in comparing long timestamp.",e);
              }
             
       }
      

       public static int compare(Text a , Text b){
              return a.compareTo(b);
       }
      
       public static int compare(LongWritable a , LongWritable b){
              return a.compareTo(b);
       }
      
       @Override
       public boolean equals(Object o) {
              if (o instanceof UserWritable) {
                     UserWritable tp = (UserWritable) o;
                     return sessionID.equals(tp.sessionID) && timestamp.equals(tp.timestamp);
              }
              return false;
       }
}

Now let us write InputFormat class with LongWritable key and UserWritable as value. Also, implement RecordReader to read file line by line.

package hive;

import java.io.IOException;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FSDataInputStream;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.fs.Seekable;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapred.FileInputFormat;
import org.apache.hadoop.mapred.FileSplit;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapred.JobConfigurable;
import org.apache.hadoop.mapred.RecordReader;
import org.apache.hadoop.mapred.Reporter;

public class UWInputFormat extends FileInputFormat<LongWritable, UserWritable> implements JobConfigurable{

       JobConf conf;
      
       @Override
       public void configure(JobConf conf) {
              this.conf = conf;         
       }
      
       @Override
       public org.apache.hadoop.mapred.RecordReader<LongWritable, UserWritable> getRecordReader(
                     org.apache.hadoop.mapred.InputSplit paramInputSplit,
                     JobConf paramJobConf, Reporter paramReporter) throws IOException {
              paramReporter.setStatus(paramInputSplit.toString());
              return new UWRecordReader(paramJobConf, (FileSplit) paramInputSplit);
       }

      
        public static class UWRecordReader implements RecordReader<LongWritable,UserWritable> {
             
              private FileSplit fileSplit;
              private Configuration conf;
              private boolean processed =false;
              private FSDataInputStream fileIn;
              private long pos;
             
             
             
             
              public UWRecordReader(Configuration job,
                   FileSplit split) throws IOException {
                     this.fileSplit =  split;
                     this.conf  = job;
                     Path file = fileSplit.getPath();

                     FileSystem fs = file.getFileSystem(conf);
                     this.fileIn = fs.open(file);
              }

             
              @Override
              public LongWritable createKey() {
                    
                     return new LongWritable();
              }
             
              @Override
              public UserWritable createValue() {
                     // TODO Auto-generated method stub
                     return new UserWritable();
              }
             
              @Override
              public long getPos() throws IOException {
                     return pos;
              }
             
              @Override
              public float getProgress() throws IOException {
                     return processed ? 1.0f : 0.0f;
              }
             
              @Override
              public boolean next(LongWritable paramK, UserWritable paramV)
                           throws IOException {
                     if(!processed){
                           pos = fileIn.getPos();
                           paramK.set(pos);
                          
                           String val = fileIn.readLine();
                          
                           if(val==null || val.trim().equals("")){
                                  processed = true;
                                  return false;
                           }
                          
                           String [] arr = val.split(",");
                           paramV.set(new Text(arr[0]), new Text(arr[1]));
                           return true;
                     }
                     return false;
              }


              @Override
              public void close() throws IOException {
                     fileIn.close();                  
              }

Now let us write Serializer-Deserializer. In below example I have only covered deserialize option which will get UserWritable object for each line read by RecordReader. 

package hive;

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

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hive.ql.io.IgnoreKeyTextOutputFormat;
import org.apache.hadoop.hive.serde.Constants;
import org.apache.hadoop.hive.serde2.SerDe;
import org.apache.hadoop.hive.serde2.SerDeException;
import org.apache.hadoop.hive.serde2.SerDeStats;
import org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector;
import org.apache.hadoop.hive.serde2.objectinspector.ObjectInspectorFactory;
import org.apache.hadoop.hive.serde2.objectinspector.StructObjectInspector;
import org.apache.hadoop.hive.serde2.typeinfo.TypeInfo;
import org.apache.hadoop.hive.serde2.typeinfo.TypeInfoUtils;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.io.Writable;

public class UWSerde implements SerDe{

       private int numColumns;
       private List<String> columnNames;
       private ArrayList<Object> row;
       private StructObjectInspector rowOI;
       private List<TypeInfo> columnTypes;
      
       @Override
       public Object deserialize(Writable arg0) throws SerDeException {
              UserWritable rowKey = (UserWritable) arg0;
              // Loop over columns in table and set values
              String colName;
              Object value;
             
              for (int c = 0; c < numColumns; c++) {
                     colName = columnNames.get(c);
                     TypeInfo ti = columnTypes.get(c);
                     if(colName.contains("session"))
                           row.set(c, rowKey.getSessionID().toString());
                     else
                           row.set(c, rowKey.getTimestamp().toString());
              }
              return row;
       }

       @Override
       public ObjectInspector getObjectInspector() throws SerDeException {
              return rowOI;
       }

       @Override
       public SerDeStats getSerDeStats() {
              // TODO Auto-generated method stub
              return null;
       }

       @Override
       public void initialize(Configuration sysProps, Properties tblProps)
                     throws SerDeException {
              String columnNameProperty = tblProps.getProperty(Constants.LIST_COLUMNS);
              columnNames = Arrays.asList(columnNameProperty.split(","));
             
              String columnTypeProperty = tblProps.getProperty(Constants.LIST_COLUMN_TYPES);
              columnTypes = TypeInfoUtils.getTypeInfosFromTypeString(columnTypeProperty);
             
              assert columnNames.size() == columnTypes.size();
              numColumns = columnNames.size();
             
              List<ObjectInspector> columnOIs = new ArrayList<ObjectInspector>(
                           columnNames.size());
              ObjectInspector oi;
             
              for (int c = 0; c < numColumns; c++) {
                     oi = TypeInfoUtils.getStandardJavaObjectInspectorFromTypeInfo(columnTypes.get(c));
                     columnOIs.add(oi);
              }
              rowOI = ObjectInspectorFactory.getStandardStructObjectInspector(columnNames, columnOIs);
              // Create an empty row object to be reused during deserialization
              row = new ArrayList<Object>(numColumns);
              for (int c = 0; c < numColumns; c++) {
                     row.add(null);
              }
             
       }

       @Override
       public Class<? extends Writable> getSerializedClass() {
              return Text.class;
       }

       @Override
       public Writable serialize(Object arg0, ObjectInspector arg1)
                     throws SerDeException {
              // TODO Auto-generated method stub
              return null;
       }

      
}

Create Jar of above code named as "hadoop-examples.jar" and open hive cli.

hive> add jar /root/hadoop-examples.jar;

hive> create table uw(
    > sessionid string,
    > timestamp string )
    > ROW FORMAT SERDE 'hive.UWSerde'
    > stored as
    > inputformat 'hive.UWInputFormat'
    > outputformat 'org.apache.hadoop.hive.ql.io.IgnoreKeyTextOutputFormat'
    > ;
OK
Time taken: 9.5 seconds

hive> load data local inpath '/root/uwserde' into table uw;
Copying data from file:/root/uwserde
Copying file: file:/root/uwserde
Loading data to table default.uw
Table default.uw stats: [numFiles=1, numRows=0, totalSize=40, rawDataSize=0]
OK
Time taken: 6.565 seconds

hive> select * from uw;
OK
kiju1233        1234567890
huhuhuhu        1233330987
Time taken: 2.689 seconds, Fetched: 2 row(s)


Comments

Post a Comment

Popular posts

Read from a hive table and write back to it using spark sql

By Dinesh
In context to Spark 2.2 - if we read from an hive table and write to same, we get following exception- scala > dy . write . mode ( "overwrite" ). insertInto ( "incremental.test2" ) org . apache . spark . sql . AnalysisException : Cannot insert overwrite into table that is also being read from .; org . apache . spark . sql . AnalysisException : Cannot insert overwrite into table that is also being read from .; 1. This error means that our process is reading from same table and writing to same table. 2. Normally, this should work as process writes to directory .hiveStaging... 3. This error occurs in case of saveAsTable method, as it overwrites entire table instead of individual partitions. 4. This error should not occur with insertInto method, as it overwrites partitions not the table. 5. A reason why this happening is because Hive table has following Spark TBLProperties in its definition. This problem will solve for insertInto met
Post a Comment
Read more




Hive Parse JSON with Array Columns and Explode it in to Multiple rows.

By Dinesh
 Say we have a JSON String like below -  { "billingCountry":"US" "orderItems":[       {          "itemId":1,          "product":"D1"       },   {          "itemId":2,          "product":"D2"       }    ] } And, our aim is to get output parsed like below -  itemId product 1 D1 2 D2   First, We can parse JSON as follows to get JSON String get_json_object(value, '$.orderItems.itemId') as itemId get_json_object(value, '$.orderItems.product') as product Second, Above will result String value like "[1,2]". We want to convert it to Array as follows - split(regexp_extract(get_json_object(value, '$.orderItems.itemId'),'^\\["(.*)\\"]$',1),'","') as itemId split(regexp_extract(get_json_object(value, '$.orderItems.product'),'^\\["(.*)\\"]$',1),&
Post a Comment
Read more




Caused by: java.lang.UnsupportedOperationException: org.apache.parquet.column.values.dictionary.PlainValuesDictionary$PlainIntegerDictionary

By Dinesh
Exception -  Caused by: java.lang.UnsupportedOperationException: org.apache.parquet.column.values.dictionary.PlainValuesDictionary$PlainIntegerDictionary at org.apache.parquet.column.Dictionary.decodeToBinary(Dictionary.java:44) at org.apache.spark.sql.execution.vectorized.ColumnVector.getUTF8String(ColumnVector.java:645) at org.apache.spark.sql.catalyst.expressions.GeneratedClass$GeneratedIterator.processNext(Unknown Source) Analysis - This might occur because of data type mismatch between Hive Table & written Parquet file. Solution - Correct the data type to match between Hive Table & Parquet
Post a Comment
Read more




org.apache.spark.sql.AnalysisException: Cannot overwrite a path that is also being read from.;

By Dinesh
  Caused by: org.apache.spark.sql.AnalysisException: Cannot overwrite a path that is also being read from.; at org.apache.spark.sql.execution.command.DDLUtils$.verifyNotReadPath(ddl.scala:906) at org.apache.spark.sql.execution.datasources.DataSourceAnalysis$$anonfun$apply$1.applyOrElse(DataSourceStrategy.scala:192) at org.apache.spark.sql.execution.datasources.DataSourceAnalysis$$anonfun$apply$1.applyOrElse(DataSourceStrategy.scala:134) at org.apache.spark.sql.catalyst.trees.TreeNode$$anonfun$2.apply(TreeNode.scala:267) at org.apache.spark.sql.catalyst.trees.TreeNode$$anonfun$2.apply(TreeNode.scala:267) at org.apache.spark.sql.catalyst.trees.CurrentOrigin$.withOrigin(TreeNode.scala:70) at org.apache.spark.sql.catalyst.trees.TreeNode.transformDown(TreeNode.scala:266) at org.apache.spark.sql.catalyst.trees.TreeNode.transform(TreeNode.scala:256) at org.apache.spark.sql.execution.datasources.DataSourceAnalysis.apply(DataSourceStrategy.scala:134) at org.apache.spark.sql.execution.datasource
Post a Comment
Read more




Hadoop Distcp Error Duplicate files in input path

By Dinesh
  One may face following error while copying data from one cluster to other, using Distcp  Command: hadoop distcp -i {src} {tgt} Error: org.apache.hadoop.toolsCopyListing$DulicateFileException: File would cause duplicates. Ideally there can't be same file names. So, what might be happening in your case is you trying to copy partitioned table from one cluster to other. And, 2 different named partitions have same file name. Your solution is to correct Source path  {src}  in your command, such that you provide path uptil partitioned sub directory, not the file. For ex - Refer below : /a/partcol=1/file1.txt /a/partcol=2/file1.txt If you use  {src}  as  "/a/*/*"  then you will get the error  "File would cause duplicates." But, if you use  {src}  as  "/a"  then you will not get error in copying.
Post a Comment
Read more