Added line end delimiter and null string support
parent
f315c1ae4b
commit
8daf918fc8
|
@ -1,15 +1,7 @@
|
||||||
package com.bizo.hive.serde.csv;
|
package com.bizo.hive.serde.csv;
|
||||||
|
|
||||||
import java.io.CharArrayReader;
|
import au.com.bytecode.opencsv.CSVReader;
|
||||||
import java.io.IOException;
|
import au.com.bytecode.opencsv.CSVWriter;
|
||||||
import java.io.Reader;
|
|
||||||
import java.io.StringWriter;
|
|
||||||
import java.io.Writer;
|
|
||||||
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.conf.Configuration;
|
||||||
import org.apache.hadoop.hive.serde.Constants;
|
import org.apache.hadoop.hive.serde.Constants;
|
||||||
import org.apache.hadoop.hive.serde2.SerDe;
|
import org.apache.hadoop.hive.serde2.SerDe;
|
||||||
|
@ -26,68 +18,74 @@ import org.apache.hadoop.hive.serde2.typeinfo.TypeInfoUtils;
|
||||||
import org.apache.hadoop.io.Text;
|
import org.apache.hadoop.io.Text;
|
||||||
import org.apache.hadoop.io.Writable;
|
import org.apache.hadoop.io.Writable;
|
||||||
|
|
||||||
import au.com.bytecode.opencsv.CSVReader;
|
import java.io.*;
|
||||||
import au.com.bytecode.opencsv.CSVWriter;
|
import java.util.ArrayList;
|
||||||
|
import java.util.Arrays;
|
||||||
|
import java.util.List;
|
||||||
|
import java.util.Properties;
|
||||||
|
|
||||||
|
|
||||||
/**
|
/**
|
||||||
* CSVSerde uses opencsv (http://opencsv.sourceforge.net/) to serialize/deserialize columns as CSV.
|
* CSVSerde uses opencsv (http://opencsv.sourceforge.net/) to serialize/deserialize columns as CSV.
|
||||||
*
|
*
|
||||||
* @author Larry Ogrodnek <ogrodnek@gmail.com>
|
* @author Larry Ogrodnek <ogrodnek@gmail.com>
|
||||||
*/
|
*/
|
||||||
public final class CSVSerde implements SerDe {
|
public final class CSVSerde implements SerDe {
|
||||||
|
|
||||||
private ObjectInspector inspector;
|
private ObjectInspector inspector;
|
||||||
private String[] outputFields;
|
private String[] outputFields;
|
||||||
private int numCols;
|
private int numCols;
|
||||||
private List<String> row;
|
private List<String> row;
|
||||||
|
|
||||||
private char separatorChar;
|
private char separatorChar;
|
||||||
private char quoteChar;
|
private char quoteChar;
|
||||||
private char escapeChar;
|
private char escapeChar;
|
||||||
|
private String lineEnd;
|
||||||
|
private String nullDefinedAs;
|
||||||
|
|
||||||
@Override
|
@Override
|
||||||
public void initialize(final Configuration conf, final Properties tbl) throws SerDeException {
|
public void initialize(final Configuration conf, final Properties tbl) throws SerDeException {
|
||||||
final List<String> columnNames = Arrays.asList(tbl.getProperty(Constants.LIST_COLUMNS).split(","));
|
final List<String> columnNames = Arrays.asList(tbl.getProperty(Constants.LIST_COLUMNS).split(","));
|
||||||
final List<TypeInfo> columnTypes = TypeInfoUtils.getTypeInfosFromTypeString(tbl.getProperty(Constants.LIST_COLUMN_TYPES));
|
final List<TypeInfo> columnTypes = TypeInfoUtils.getTypeInfosFromTypeString(tbl.getProperty(Constants.LIST_COLUMN_TYPES));
|
||||||
|
|
||||||
numCols = columnNames.size();
|
numCols = columnNames.size();
|
||||||
|
|
||||||
final List<ObjectInspector> columnOIs = new ArrayList<ObjectInspector>(numCols);
|
final List<ObjectInspector> columnOIs = new ArrayList<ObjectInspector>(numCols);
|
||||||
|
|
||||||
for (int i=0; i< numCols; i++) {
|
for (int i=0; i< numCols; i++) {
|
||||||
columnOIs.add(PrimitiveObjectInspectorFactory.javaStringObjectInspector);
|
columnOIs.add(PrimitiveObjectInspectorFactory.javaStringObjectInspector);
|
||||||
}
|
}
|
||||||
|
|
||||||
this.inspector = ObjectInspectorFactory.getStandardStructObjectInspector(columnNames, columnOIs);
|
this.inspector = ObjectInspectorFactory.getStandardStructObjectInspector(columnNames, columnOIs);
|
||||||
this.outputFields = new String[numCols];
|
this.outputFields = new String[numCols];
|
||||||
row = new ArrayList<String>(numCols);
|
row = new ArrayList<String>(numCols);
|
||||||
|
|
||||||
for (int i=0; i< numCols; i++) {
|
for (int i=0; i< numCols; i++) {
|
||||||
row.add(null);
|
row.add(null);
|
||||||
}
|
}
|
||||||
|
|
||||||
separatorChar = getProperty(tbl, "separatorChar", CSVWriter.DEFAULT_SEPARATOR);
|
separatorChar = getProperty(tbl, "separatorChar", CSVWriter.DEFAULT_SEPARATOR);
|
||||||
quoteChar = getProperty(tbl, "quoteChar", CSVWriter.DEFAULT_QUOTE_CHARACTER);
|
quoteChar = getProperty(tbl, "quoteChar", CSVWriter.DEFAULT_QUOTE_CHARACTER);
|
||||||
escapeChar = getProperty(tbl, "escapeChar", CSVWriter.DEFAULT_ESCAPE_CHARACTER);
|
escapeChar = getProperty(tbl, "escapeChar", CSVWriter.DEFAULT_ESCAPE_CHARACTER);
|
||||||
|
lineEnd = tbl.getProperty("lineEnd", CSVWriter.DEFAULT_LINE_END);
|
||||||
|
nullDefinedAs = tbl.getProperty("nullDefinedAs", "\\N");
|
||||||
}
|
}
|
||||||
|
|
||||||
private final char getProperty(final Properties tbl, final String property, final char def) {
|
private final char getProperty(final Properties tbl, final String property, final char def) {
|
||||||
final String val = tbl.getProperty(property);
|
final String val = tbl.getProperty(property);
|
||||||
|
|
||||||
if (val != null) {
|
if (val != null) {
|
||||||
return val.charAt(0);
|
return val.charAt(0);
|
||||||
}
|
}
|
||||||
|
|
||||||
return def;
|
return def;
|
||||||
}
|
}
|
||||||
|
|
||||||
@Override
|
@Override
|
||||||
public Writable serialize(Object obj, ObjectInspector objInspector) throws SerDeException {
|
public Writable serialize(Object obj, ObjectInspector objInspector) throws SerDeException {
|
||||||
final StructObjectInspector outputRowOI = (StructObjectInspector) objInspector;
|
final StructObjectInspector outputRowOI = (StructObjectInspector) objInspector;
|
||||||
final List<? extends StructField> outputFieldRefs = outputRowOI.getAllStructFieldRefs();
|
final List<? extends StructField> outputFieldRefs = outputRowOI.getAllStructFieldRefs();
|
||||||
|
|
||||||
if (outputFieldRefs.size() != numCols) {
|
if (outputFieldRefs.size() != numCols) {
|
||||||
throw new SerDeException("Cannot serialize the object because there are "
|
throw new SerDeException("Cannot serialize the object because there are "
|
||||||
+ outputFieldRefs.size() + " fields but the table has " + numCols + " columns.");
|
+ outputFieldRefs.size() + " fields but the table has " + numCols + " columns.");
|
||||||
|
@ -97,45 +95,48 @@ public final class CSVSerde implements SerDe {
|
||||||
for (int c = 0; c < numCols; c++) {
|
for (int c = 0; c < numCols; c++) {
|
||||||
final Object field = outputRowOI.getStructFieldData(obj, outputFieldRefs.get(c));
|
final Object field = outputRowOI.getStructFieldData(obj, outputFieldRefs.get(c));
|
||||||
final ObjectInspector fieldOI = outputFieldRefs.get(c).getFieldObjectInspector();
|
final ObjectInspector fieldOI = outputFieldRefs.get(c).getFieldObjectInspector();
|
||||||
|
|
||||||
// The data must be of type String
|
// The data must be of type String
|
||||||
final StringObjectInspector fieldStringOI = (StringObjectInspector) fieldOI;
|
final StringObjectInspector fieldStringOI = (StringObjectInspector) fieldOI;
|
||||||
|
|
||||||
// Convert the field to Java class String, because objects of String type
|
// Convert the field to Java class String, because objects of String type
|
||||||
// can be stored in String, Text, or some other classes.
|
// can be stored in String, Text, or some other classes.
|
||||||
outputFields[c] = fieldStringOI.getPrimitiveJavaObject(field);
|
outputFields[c] = fieldStringOI.getPrimitiveJavaObject(field);
|
||||||
|
if (outputFields[c] == null) {
|
||||||
|
outputFields[c] = nullDefinedAs;
|
||||||
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
final StringWriter writer = new StringWriter();
|
final StringWriter writer = new StringWriter();
|
||||||
final CSVWriter csv = newWriter(writer, separatorChar, quoteChar, escapeChar);
|
final CSVWriter csv = newWriter(writer, separatorChar, quoteChar, escapeChar, lineEnd);
|
||||||
|
|
||||||
try {
|
try {
|
||||||
csv.writeNext(outputFields);
|
csv.writeNext(outputFields);
|
||||||
csv.close();
|
csv.close();
|
||||||
|
|
||||||
return new Text(writer.toString());
|
return new Text(writer.toString());
|
||||||
} catch (final IOException ioe) {
|
} catch (final IOException ioe) {
|
||||||
throw new SerDeException(ioe);
|
throw new SerDeException(ioe);
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
@Override
|
@Override
|
||||||
public Object deserialize(final Writable blob) throws SerDeException {
|
public Object deserialize(final Writable blob) throws SerDeException {
|
||||||
Text rowText = (Text) blob;
|
Text rowText = (Text) blob;
|
||||||
|
|
||||||
CSVReader csv = null;
|
CSVReader csv = null;
|
||||||
try {
|
try {
|
||||||
csv = newReader(new CharArrayReader(rowText.toString().toCharArray()), separatorChar, quoteChar, escapeChar);
|
csv = newReader(new CharArrayReader(rowText.toString().toCharArray()), separatorChar, quoteChar, escapeChar);
|
||||||
final String[] read = csv.readNext();
|
final String[] read = csv.readNext();
|
||||||
|
|
||||||
for (int i=0; i< numCols; i++) {
|
for (int i=0; i< numCols; i++) {
|
||||||
if (read != null && i < read.length) {
|
if (read != null && i < read.length) {
|
||||||
row.set(i, read[i]);
|
row.set(i, read[i] == nullDefinedAs ? null : read[i]);
|
||||||
} else {
|
} else {
|
||||||
row.set(i, null);
|
row.set(i, null);
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
return row;
|
return row;
|
||||||
} catch (final Exception e) {
|
} catch (final Exception e) {
|
||||||
throw new SerDeException(e);
|
throw new SerDeException(e);
|
||||||
|
@ -149,22 +150,22 @@ public final class CSVSerde implements SerDe {
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
private CSVReader newReader(final Reader reader, char separator, char quote, char escape) {
|
private static CSVReader newReader(final Reader reader, char separator, char quote, char escape) {
|
||||||
// CSVReader will throw an exception if any of separator, quote, or escape is the same, but
|
// CSVReader will throw an exception if any of separator, quote, or escape is the same, but
|
||||||
// the CSV format specifies that the escape character and quote char are the same... very weird
|
// the CSV format specifies that the escape character and quote char are the same... very weird
|
||||||
if (CSVWriter.DEFAULT_ESCAPE_CHARACTER == escape) {
|
if (CSVWriter.DEFAULT_ESCAPE_CHARACTER == escape) {
|
||||||
return new CSVReader(reader, separator, quote);
|
return new CSVReader(reader, separator, quote);
|
||||||
} else {
|
} else {
|
||||||
return new CSVReader(reader, separator, quote, escape);
|
return new CSVReader(reader, separator, quote, escape);
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
private CSVWriter newWriter(final Writer writer, char separator, char quote, char escape) {
|
private static CSVWriter newWriter(final Writer writer, char separator, char quote, char escape, String lineEnd) {
|
||||||
if (CSVWriter.DEFAULT_ESCAPE_CHARACTER == escape) {
|
if (CSVWriter.DEFAULT_ESCAPE_CHARACTER == escape) {
|
||||||
return new CSVWriter(writer, separator, quote, "");
|
return new CSVWriter(writer, separator, quote, lineEnd);
|
||||||
} else {
|
} else {
|
||||||
return new CSVWriter(writer, separator, quote, escape, "");
|
return new CSVWriter(writer, separator, quote, escape, lineEnd);
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -177,7 +178,7 @@ public final class CSVSerde implements SerDe {
|
||||||
public Class<? extends Writable> getSerializedClass() {
|
public Class<? extends Writable> getSerializedClass() {
|
||||||
return Text.class;
|
return Text.class;
|
||||||
}
|
}
|
||||||
|
|
||||||
public SerDeStats getSerDeStats() {
|
public SerDeStats getSerDeStats() {
|
||||||
return null;
|
return null;
|
||||||
}
|
}
|
||||||
|
|
Loading…
Reference in New Issue