use of org.apache.hadoop.hive.common.type.HiveChar in project hive by apache.
the class AvroDeserializer method deserializePrimitive.
private Object deserializePrimitive(Object datum, Schema fileSchema, Schema recordSchema, PrimitiveTypeInfo columnType) throws AvroSerdeException {
switch(columnType.getPrimitiveCategory()) {
case STRING:
// To workaround AvroUTF8
return datum.toString();
// and convert it to a string. Yay!
case BINARY:
if (recordSchema.getType() == Type.FIXED) {
Fixed fixed = (Fixed) datum;
return fixed.bytes();
} else if (recordSchema.getType() == Type.BYTES) {
return AvroSerdeUtils.getBytesFromByteBuffer((ByteBuffer) datum);
} else {
throw new AvroSerdeException("Unexpected Avro schema for Binary TypeInfo: " + recordSchema.getType());
}
case DECIMAL:
if (fileSchema == null) {
throw new AvroSerdeException("File schema is missing for decimal field. Reader schema is " + columnType);
}
int scale = 0;
try {
scale = AvroSerdeUtils.getIntFromSchema(fileSchema, AvroSerDe.AVRO_PROP_SCALE);
} catch (Exception ex) {
throw new AvroSerdeException("Failed to obtain scale value from file schema: " + fileSchema, ex);
}
HiveDecimal dec = AvroSerdeUtils.getHiveDecimalFromByteBuffer((ByteBuffer) datum, scale);
JavaHiveDecimalObjectInspector oi = (JavaHiveDecimalObjectInspector) PrimitiveObjectInspectorFactory.getPrimitiveJavaObjectInspector((DecimalTypeInfo) columnType);
return oi.set(null, dec);
case CHAR:
if (fileSchema == null) {
throw new AvroSerdeException("File schema is missing for char field. Reader schema is " + columnType);
}
int maxLength = 0;
try {
maxLength = AvroSerdeUtils.getIntFromSchema(fileSchema, AvroSerDe.AVRO_PROP_MAX_LENGTH);
} catch (Exception ex) {
throw new AvroSerdeException("Failed to obtain maxLength value for char field from file schema: " + fileSchema, ex);
}
String str = datum.toString();
HiveChar hc = new HiveChar(str, maxLength);
return hc;
case VARCHAR:
if (fileSchema == null) {
throw new AvroSerdeException("File schema is missing for varchar field. Reader schema is " + columnType);
}
maxLength = 0;
try {
maxLength = AvroSerdeUtils.getIntFromSchema(fileSchema, AvroSerDe.AVRO_PROP_MAX_LENGTH);
} catch (Exception ex) {
throw new AvroSerdeException("Failed to obtain maxLength value for varchar field from file schema: " + fileSchema, ex);
}
str = datum.toString();
HiveVarchar hvc = new HiveVarchar(str, maxLength);
return hvc;
case DATE:
{
if (recordSchema.getType() != Type.INT) {
throw new AvroSerdeException("Unexpected Avro schema for Date TypeInfo: " + recordSchema.getType());
}
final boolean skipProlepticConversion;
if (writerProleptic != null) {
skipProlepticConversion = writerProleptic;
} else {
if (configuration != null) {
skipProlepticConversion = HiveConf.getBoolVar(configuration, HiveConf.ConfVars.HIVE_AVRO_PROLEPTIC_GREGORIAN_DEFAULT);
} else {
skipProlepticConversion = HiveConf.ConfVars.HIVE_AVRO_PROLEPTIC_GREGORIAN_DEFAULT.defaultBoolVal;
}
}
return Date.ofEpochMilli(DateWritableV2.daysToMillis(skipProlepticConversion ? (Integer) datum : CalendarUtils.convertDateToProleptic((Integer) datum)));
}
case TIMESTAMP:
{
if (recordSchema.getType() != Type.LONG) {
throw new AvroSerdeException("Unexpected Avro schema for Date TypeInfo: " + recordSchema.getType());
}
// If a time zone is found in file metadata (property name: writer.time.zone), convert the
// timestamp to that (writer) time zone in order to emulate time zone agnostic behavior.
// If not, then the file was written by an older version of hive, so we convert the timestamp
// to the server's (reader) time zone for backwards compatibility reasons - unless the
// session level configuration hive.avro.timestamp.skip.conversion is set to true, in which
// case we assume it was written by a time zone agnostic writer, so we don't convert it.
final boolean skipUTCConversion;
if (configuration != null) {
skipUTCConversion = HiveConf.getBoolVar(configuration, HiveConf.ConfVars.HIVE_AVRO_TIMESTAMP_SKIP_CONVERSION);
} else {
skipUTCConversion = HiveConf.ConfVars.HIVE_AVRO_TIMESTAMP_SKIP_CONVERSION.defaultBoolVal;
}
final boolean legacyConversion;
if (writerZoneConversionLegacy != null) {
legacyConversion = writerZoneConversionLegacy;
} else if (writerTimezone != null) {
legacyConversion = false;
} else if (configuration != null) {
legacyConversion = HiveConf.getBoolVar(configuration, HiveConf.ConfVars.HIVE_AVRO_TIMESTAMP_LEGACY_CONVERSION_ENABLED);
} else {
legacyConversion = HiveConf.ConfVars.HIVE_AVRO_TIMESTAMP_LEGACY_CONVERSION_ENABLED.defaultBoolVal;
}
ZoneId convertToTimeZone;
if (writerTimezone != null) {
convertToTimeZone = writerTimezone;
} else if (skipUTCConversion) {
convertToTimeZone = ZoneOffset.UTC;
} else {
convertToTimeZone = TimeZone.getDefault().toZoneId();
}
final boolean skipProlepticConversion;
if (writerProleptic != null) {
skipProlepticConversion = writerProleptic;
} else {
if (configuration != null) {
skipProlepticConversion = HiveConf.getBoolVar(configuration, HiveConf.ConfVars.HIVE_AVRO_PROLEPTIC_GREGORIAN_DEFAULT);
} else {
skipProlepticConversion = HiveConf.ConfVars.HIVE_AVRO_PROLEPTIC_GREGORIAN_DEFAULT.defaultBoolVal;
}
}
Timestamp timestamp = TimestampTZUtil.convertTimestampToZone(Timestamp.ofEpochMilli((Long) datum), ZoneOffset.UTC, convertToTimeZone, legacyConversion);
if (!skipProlepticConversion) {
timestamp = Timestamp.ofEpochMilli(CalendarUtils.convertTimeToProleptic(timestamp.toEpochMilli()));
}
return timestamp;
}
default:
return datum;
}
}
use of org.apache.hadoop.hive.common.type.HiveChar in project hive by apache.
the class TestSearchArgumentImpl method testBuilderComplexTypes2.
@Test
public void testBuilderComplexTypes2() throws Exception {
SearchArgumentImpl sarg = (SearchArgumentImpl) SearchArgumentFactory.newBuilder().startAnd().lessThan("x", PredicateLeaf.Type.DATE, Date.valueOf("2005-3-12")).lessThanEquals("y", PredicateLeaf.Type.STRING, new HiveChar("hi", 10).toString()).equals("z", PredicateLeaf.Type.DECIMAL, new HiveDecimalWritable("1.0")).end().build();
assertEquals("leaf-0 = (LESS_THAN x 2005-03-12), " + "leaf-1 = (LESS_THAN_EQUALS y hi ), " + "leaf-2 = (EQUALS z 1), " + "expr = (and leaf-0 leaf-1 leaf-2)", sarg.toOldString());
sarg = (SearchArgumentImpl) SearchArgumentFactory.newBuilder().startNot().startOr().isNull("x", PredicateLeaf.Type.LONG).between("y", PredicateLeaf.Type.DECIMAL, new HiveDecimalWritable("10"), new HiveDecimalWritable("20.0")).in("z", PredicateLeaf.Type.LONG, 1L, 2L, 3L).nullSafeEquals("a", PredicateLeaf.Type.STRING, new HiveVarchar("stinger", 100).toString()).end().end().build();
assertEquals("leaf-0 = (IS_NULL x), " + "leaf-1 = (BETWEEN y 10 20), " + "leaf-2 = (IN z 1 2 3), " + "leaf-3 = (NULL_SAFE_EQUALS a stinger), " + "expr = (and (not leaf-0) (not leaf-1) (not leaf-2) (not leaf-3))", sarg.toOldString());
}
use of org.apache.hadoop.hive.common.type.HiveChar in project hive by apache.
the class TestSearchArgumentImpl method testBuilderComplexTypes.
@Test
public void testBuilderComplexTypes() throws Exception {
SearchArgumentImpl sarg = (SearchArgumentImpl) SearchArgumentFactory.newBuilder().startAnd().lessThan("x", PredicateLeaf.Type.DATE, Date.valueOf("1970-1-11")).lessThanEquals("y", PredicateLeaf.Type.STRING, new HiveChar("hi", 10).toString()).equals("z", PredicateLeaf.Type.DECIMAL, new HiveDecimalWritable("1.0")).end().build();
assertEquals("leaf-0 = (LESS_THAN x 1970-01-11), " + "leaf-1 = (LESS_THAN_EQUALS y hi ), " + "leaf-2 = (EQUALS z 1), " + "expr = (and leaf-0 leaf-1 leaf-2)", sarg.toOldString());
sarg = (SearchArgumentImpl) SearchArgumentFactory.newBuilder().startNot().startOr().isNull("x", PredicateLeaf.Type.LONG).between("y", PredicateLeaf.Type.DECIMAL, new HiveDecimalWritable("10"), new HiveDecimalWritable("20.0")).in("z", PredicateLeaf.Type.LONG, 1L, 2L, 3L).nullSafeEquals("a", PredicateLeaf.Type.STRING, new HiveVarchar("stinger", 100).toString()).end().end().build();
assertEquals("leaf-0 = (IS_NULL x), " + "leaf-1 = (BETWEEN y 10 20), " + "leaf-2 = (IN z 1 2 3), " + "leaf-3 = (NULL_SAFE_EQUALS a stinger), " + "expr = (and (not leaf-0) (not leaf-1) (not leaf-2) (not leaf-3))", sarg.toOldString());
}
use of org.apache.hadoop.hive.common.type.HiveChar in project hive by apache.
the class TestParquetFilterPredicate method testFilterCharColumnGreaterThan.
@Test
public void testFilterCharColumnGreaterThan() throws Exception {
SearchArgument sarg = SearchArgumentFactory.newBuilder().startNot().lessThanEquals("a", PredicateLeaf.Type.STRING, new HiveChar("apple", 10).toString()).end().build();
MessageType schema = MessageTypeParser.parseMessageType("message test {required binary a;}");
Map<String, TypeInfo> columnTypes = new HashMap<>();
columnTypes.put("a", TypeInfoFactory.getCharTypeInfo(10));
FilterPredicate p = ParquetFilterPredicateConverter.toFilterPredicate(sarg, schema, columnTypes);
String expected = "not(lteq(a, Binary{\"apple\"}))";
assertEquals(expected, p.toString());
}
use of org.apache.hadoop.hive.common.type.HiveChar in project hive by apache.
the class TestParquetFilterPredicate method testFilterCharColumnBetween.
@Test
public void testFilterCharColumnBetween() throws Exception {
SearchArgument sarg = SearchArgumentFactory.newBuilder().between("a", PredicateLeaf.Type.STRING, new HiveChar("apple", 10).toString(), new HiveChar("pear", 10).toString()).build();
MessageType schema = MessageTypeParser.parseMessageType("message test {required binary a;}");
Map<String, TypeInfo> columnTypes = new HashMap<>();
columnTypes.put("a", TypeInfoFactory.getCharTypeInfo(10));
FilterPredicate p = ParquetFilterPredicateConverter.toFilterPredicate(sarg, schema, columnTypes);
String expected = "and(lteq(a, Binary{\"pear\"}), not(lt(a, Binary{\"apple\"})))";
assertEquals(expected, p.toString());
}
Aggregations