diff --git itests/qtest/pom.xml itests/qtest/pom.xml index 1a19610..919036f 100644 --- itests/qtest/pom.xml +++ itests/qtest/pom.xml @@ -38,7 +38,7 @@ false stats_counter_partitioned.q,list_bucket_dml_10.q,input16_cc.q,scriptfile1.q,scriptfile1_win.q,bucket4.q,bucketmapjoin6.q,disable_merge_for_bucketing.q,reduce_deduplicate.q,smb_mapjoin_8.q,join1.q,groupby2.q,bucketizedhiveinputformat.q,bucketmapjoin7.q,optrstat_groupby.q,bucket_num_reducers.q,bucket5.q,load_fs2.q,bucket_num_reducers2.q,infer_bucket_sort_merge.q,infer_bucket_sort_reducers_power_two.q,infer_bucket_sort_dyn_part.q,infer_bucket_sort_bucketed_table.q,infer_bucket_sort_map_operators.q,infer_bucket_sort_num_buckets.q,leftsemijoin_mr.q,schemeAuthority.q,schemeAuthority2.q,truncate_column_buckets.q,remote_script.q,,load_hdfs_file_with_space_in_the_name.q,parallel_orderby.q,import_exported_table.q,stats_counter.q,auto_sortmerge_join_16.q,quotedid_smb.q,file_with_header_footer.q,external_table_with_space_in_location_path.q,root_dir_external_table.q,index_bitmap3.q,ql_rewrite_gbtoidx.q,index_bitmap_auto.q,udf_using.q cluster_tasklog_retrieval.q,minimr_broken_pipe.q,mapreduce_stack_trace.q,mapreduce_stack_trace_turnoff.q,mapreduce_stack_trace_hadoop20.q,mapreduce_stack_trace_turnoff_hadoop20.q,file_with_header_footer_negative.q,udf_local_resource.q - tez_fsstat.q,mapjoin_decimal.q,tez_join_tests.q,tez_joins_explain.q,mrr.q,tez_dml.q,tez_insert_overwrite_local_directory_1.q,tez_union.q,bucket_map_join_tez1.q,bucket_map_join_tez2.q,tez_schema_evolution.q + tez_fsstat.q,mapjoin_decimal.q,tez_join_tests.q,tez_joins_explain.q,mrr.q,tez_dml.q,tez_insert_overwrite_local_directory_1.q,tez_union.q,bucket_map_join_tez1.q,bucket_map_join_tez2.q,tez_schema_evolution.q,udf_concat_tez.q cross_product_check_1.q,cross_product_check_2.q,dynpart_sort_opt_vectorization.q,dynpart_sort_optimization.q,orc_analyze.q,join0.q,join1.q,auto_join0.q,auto_join1.q,bucket2.q,bucket3.q,bucket4.q,count.q,create_merge_compressed.q,cross_join.q,ctas.q,custom_input_output_format.q,disable_merge_for_bucketing.q,enforce_order.q,filter_join_breaktask.q,filter_join_breaktask2.q,groupby1.q,groupby2.q,groupby3.q,having.q,insert1.q,insert_into1.q,insert_into2.q,leftsemijoin.q,limit_pushdown.q,load_dyn_part1.q,load_dyn_part2.q,load_dyn_part3.q,mapjoin_mapjoin.q,mapreduce1.q,mapreduce2.q,merge1.q,merge2.q,metadata_only_queries.q,sample1.q,subquery_in.q,subquery_exists.q,vectorization_15.q,ptf.q,stats_counter.q,stats_noscan_1.q,stats_counter_partitioned.q,union2.q,union3.q,union4.q,union5.q,union6.q,union7.q,union8.q,union9.q,transform1.q,transform2.q,transform_ppr1.q,transform_ppr2.q,script_env_var1.q,script_env_var2.q,script_pipe.q,scriptfile1.q add_part_exist.q,alter1.q,alter2.q,alter4.q,alter5.q,alter_rename_partition.q,alter_rename_partition_authorization.q,archive.q,archive_corrupt.q,archive_multi.q,archive_mr_1806.q,archive_multi_mr_1806.q,authorization_1.q,authorization_2.q,authorization_4.q,authorization_5.q,authorization_6.q,authorization_7.q,ba_table1.q,ba_table2.q,ba_table3.q,ba_table_udfs.q,binary_table_bincolserde.q,binary_table_colserde.q,cluster.q,columnarserde_create_shortcut.q,combine2.q,constant_prop.q,create_nested_type.q,create_or_replace_view.q,create_struct_table.q,create_union_table.q,database.q,database_location.q,database_properties.q,ddltime.q,describe_database_json.q,drop_database_removes_partition_dirs.q,escape1.q,escape2.q,exim_00_nonpart_empty.q,exim_01_nonpart.q,exim_02_00_part_empty.q,exim_02_part.q,exim_03_nonpart_over_compat.q,exim_04_all_part.q,exim_04_evolved_parts.q,exim_05_some_part.q,exim_06_one_part.q,exim_07_all_part_over_nonoverlap.q,exim_08_nonpart_rename.q,exim_09_part_spec_nonoverlap.q,exim_10_external_managed.q,exim_11_managed_external.q,exim_12_external_location.q,exim_13_managed_location.q,exim_14_managed_location_over_existing.q,exim_15_external_part.q,exim_16_part_external.q,exim_17_part_managed.q,exim_18_part_external.q,exim_19_00_part_external_location.q,exim_19_part_external_location.q,exim_20_part_managed_location.q,exim_21_export_authsuccess.q,exim_22_import_exist_authsuccess.q,exim_23_import_part_authsuccess.q,exim_24_import_nonexist_authsuccess.q,global_limit.q,groupby_complex_types.q,groupby_complex_types_multi_single_reducer.q,index_auth.q,index_auto.q,index_auto_empty.q,index_bitmap.q,index_bitmap1.q,index_bitmap2.q,index_bitmap3.q,index_bitmap_auto.q,index_bitmap_rc.q,index_compact.q,index_compact_1.q,index_compact_2.q,index_compact_3.q,index_stale_partitioned.q,init_file.q,input16.q,input16_cc.q,input46.q,input_columnarserde.q,input_dynamicserde.q,input_lazyserde.q,input_testxpath3.q,input_testxpath4.q,insert2_overwrite_partitions.q,insertexternal1.q,join_thrift.q,lateral_view.q,load_binary_data.q,load_exist_part_authsuccess.q,load_nonpart_authsuccess.q,load_part_authsuccess.q,loadpart_err.q,lock1.q,lock2.q,lock3.q,lock4.q,merge_dynamic_partition.q,multi_insert.q,multi_insert_move_tasks_share_dependencies.q,null_column.q,ppd_clusterby.q,query_with_semi.q,rename_column.q,sample6.q,sample_islocalmode_hook.q,set_processor_namespaces.q,show_tables.q,source.q,split_sample.q,str_to_map.q,transform1.q,udaf_collect_set.q,udaf_context_ngrams.q,udaf_histogram_numeric.q,udaf_ngrams.q,udaf_percentile_approx.q,udf_array.q,udf_bitmap_and.q,udf_bitmap_or.q,udf_explode.q,udf_format_number.q,udf_map.q,udf_map_keys.q,udf_map_values.q,udf_max.q,udf_min.q,udf_named_struct.q,udf_percentile.q,udf_printf.q,udf_sentences.q,udf_sort_array.q,udf_split.q,udf_struct.q,udf_substr.q,udf_translate.q,udf_union.q,udf_xpath.q,udtf_stack.q,view.q,virtual_column.q diff --git ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorExpressionDescriptor.java ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorExpressionDescriptor.java index bfdd3ce..e6d10af 100644 --- ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorExpressionDescriptor.java +++ ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorExpressionDescriptor.java @@ -150,6 +150,10 @@ public Builder setInputExpressionType(int index, InputExpressionType type) { return this; } + public ArgumentType getArgumentType(int index) { + return argTypes[index]; + } + public Descriptor build() { return new Descriptor(mode, argCount, argTypes, exprTypes); } diff --git ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorizationContext.java ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorizationContext.java index 9772b4d..45b8538 100644 --- ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorizationContext.java +++ ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorizationContext.java @@ -39,6 +39,7 @@ import org.apache.hadoop.hive.ql.exec.FunctionInfo; import org.apache.hadoop.hive.ql.exec.FunctionRegistry; import org.apache.hadoop.hive.ql.exec.UDF; +import org.apache.hadoop.hive.ql.exec.vector.VectorExpressionDescriptor.ArgumentType; import org.apache.hadoop.hive.ql.exec.vector.VectorExpressionDescriptor.InputExpressionType; import org.apache.hadoop.hive.ql.exec.vector.VectorExpressionDescriptor.Mode; import org.apache.hadoop.hive.ql.exec.vector.expressions.*; @@ -100,6 +101,7 @@ import org.apache.hadoop.hive.serde2.typeinfo.HiveDecimalUtils; import org.apache.hadoop.hive.serde2.typeinfo.PrimitiveTypeInfo; import org.apache.hadoop.hive.serde2.typeinfo.TypeInfo; +import org.apache.hadoop.hive.serde2.typeinfo.TypeInfoFactory; /** * Context class for vectorization execution. @@ -484,7 +486,7 @@ private ExprNodeDesc getImplicitCastExpression(GenericUDF udf, ExprNodeDesc chil } else { // Casts to exact types including long to double etc. are needed in some special cases. - if (udf instanceof GenericUDFCoalesce) { + if (udf == null || udf instanceof GenericUDFCoalesce) { GenericUDF genericUdf = getGenericUDFForCast(castType); List children = new ArrayList(); children.add(child); @@ -784,20 +786,105 @@ private VectorExpression getVectorExpressionForUdf(Class udf, List vclass = this.vMap.getVectorExpressionClass(udf, descriptor); if (vclass == null) { if (LOG.isDebugEnabled()) { - LOG.debug("No vector udf found for "+udf.getSimpleName() + ", descriptor: "+descriptor); + LOG.debug("Failed to vectorizing " + udf.getSimpleName() + " with " + descriptor); + } + if (numChildren > 0) { + casters = getCaster(udf, childExpr, builder); + if (casters != null) { + if (LOG.isDebugEnabled()) { + LOG.debug("Retry vectorizing " + udf.getSimpleName() + " with " + builder.build()); + } + vclass = this.vMap.getVectorExpressionClass(udf, builder.build()); + } + } + if (vclass == null) { + if (LOG.isDebugEnabled()) { + LOG.debug("No vector udf found for " + udf.getSimpleName() + ", descriptor: " + descriptor); + } + return null; } - return null; } Mode childrenMode = getChildrenMode(mode, udf); - return createVectorExpression(vclass, childExpr, childrenMode, returnType); + return createVectorExpression(vclass, childExpr, childrenMode, casters, returnType); + } + + private TypeInfo[] getCaster(Class udf, List childExpr, + VectorExpressionDescriptor.Builder builder) { + String[] preferred = getPreferredTypes(udf, childExpr.size()); + if (preferred == null) { + return null; + } + TypeInfo[] casters = new TypeInfo[preferred.length]; + for (int i = 0; i < preferred.length; i++) { + ExprNodeDesc child = childExpr.get(i); + ArgumentType sourceType = builder.getArgumentType(i); + ArgumentType targetType = ArgumentType.getType(preferred[i]); + if (sourceType != targetType) { + try { + casters[i] = TypeInfoFactory.getPrimitiveTypeInfo(targetType.name().toLowerCase()); + builder.setArgumentType(i, targetType); + } catch (Exception e) { + LOG.info("Invalid type " + targetType.name(), e); + return null; + } + } + } + return casters; + } + + private String[] getPreferredTypes(Class udf, int numChildren) { + VectorizedExpressions annotation = udf.getAnnotation(VectorizedExpressions.class); + String[] preferred = toPreferredTypes(numChildren, annotation.preferredType()); + if (LOG.isDebugEnabled()) { + LOG.debug("Preferred arguments " + udf.getSimpleName() + Arrays.toString(preferred)); + } + return preferred; + } + + static String[] toPreferredTypes(int numChildren, String description) { + String[] spec = splitAndTrim(description); + if (spec.length - 1 > numChildren) { + return null; + } + int i = 0; + for (; i < spec.length; i++) { + if (spec[i].endsWith("...")) { + break; + } + } + if (i == spec.length && numChildren != spec.length) { + return null; + } + String[] types = new String[numChildren]; + System.arraycopy(spec, 0, types, 0, i); + System.arraycopy(spec, i + 1, types, types.length - i, spec.length - i - 1); + if (i != spec.length) { + Arrays.fill(types, i, types.length - (spec.length - i - 1), spec[i].substring(0, spec[i].length() - 3)); + } + return types; + } + + private static String[] splitAndTrim(String description) { + String[] split = description.split(","); + for (int i = 0; i < split.length; i++) { + split[i] = split[i].trim(); + } + return split; } private VectorExpression createVectorExpression(Class vectorClass, List childExpr, Mode childrenMode, TypeInfo returnType) throws HiveException { + return createVectorExpression(vectorClass, childExpr, childrenMode, null, returnType); + } + + private VectorExpression createVectorExpression(Class vectorClass, + List childExpr, Mode childrenMode, TypeInfo[] casters, TypeInfo returnType) + throws HiveException { int numChildren = childExpr == null ? 0: childExpr.size(); VectorExpression.Type [] inputTypes = new VectorExpression.Type[numChildren]; List children = new ArrayList(); @@ -805,24 +892,27 @@ private VectorExpression createVectorExpression(Class vectorClass, try { for (int i = 0; i < numChildren; i++) { ExprNodeDesc child = childExpr.get(i); - inputTypes[i] = VectorExpression.Type.getValue(child.getTypeInfo().getTypeName()); + if (casters != null && casters[i] != null) { + child = getImplicitCastExpression(null, child, casters[i]); + } if (child instanceof ExprNodeGenericFuncDesc) { VectorExpression vChild = getVectorExpression(child, childrenMode); - children.add(vChild); - arguments[i] = vChild.getOutputColumn(); + children.add(vChild); + arguments[i] = vChild.getOutputColumn(); } else if (child instanceof ExprNodeColumnDesc) { int colIndex = getInputColumnIndex((ExprNodeColumnDesc) child); - if (childrenMode == Mode.FILTER) { - // In filter mode, the column must be a boolean - children.add(new SelectColumnIsTrue(colIndex)); - } - arguments[i] = colIndex; + if (childrenMode == Mode.FILTER) { + // In filter mode, the column must be a boolean + children.add(new SelectColumnIsTrue(colIndex)); + } + arguments[i] = colIndex; } else if (child instanceof ExprNodeConstantDesc) { Object scalarValue = getVectorTypeScalarValue((ExprNodeConstantDesc) child); arguments[i] = scalarValue; } else { - throw new HiveException("Cannot handle expression type: " + child.getClass().getSimpleName()); + throw new HiveException("Cannot handle expression: " + child); } + inputTypes[i] = VectorExpression.Type.getValue(child.getTypeInfo().getTypeName()); } VectorExpression vectorExpression = instantiateExpression(vectorClass, returnType, arguments); vectorExpression.setInputTypes(inputTypes); diff --git ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorizedExpressions.java ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorizedExpressions.java index e7d1957..a6ec697 100644 --- ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorizedExpressions.java +++ ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorizedExpressions.java @@ -26,6 +26,8 @@ @Retention(RetentionPolicy.RUNTIME) public @interface VectorizedExpressions { + String preferredType() default ""; + Class[] value(); } diff --git ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFConcat.java ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFConcat.java index d4d7e7c..d51f573 100644 --- ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFConcat.java +++ ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFConcat.java @@ -27,11 +27,11 @@ import org.apache.hadoop.hive.ql.exec.vector.expressions.StringConcatColScalar; import org.apache.hadoop.hive.ql.exec.vector.expressions.StringConcatScalarCol; import org.apache.hadoop.hive.ql.metadata.HiveException; +import org.apache.hadoop.hive.serde2.objectinspector.ConstantObjectInspector; import org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector; import org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector.Category; import org.apache.hadoop.hive.serde2.objectinspector.PrimitiveObjectInspector; import org.apache.hadoop.hive.serde2.objectinspector.PrimitiveObjectInspector.PrimitiveCategory; -import org.apache.hadoop.hive.serde2.objectinspector.primitive.BinaryObjectInspector; import org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorConverter.StringConverter; import org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory; import org.apache.hadoop.hive.serde2.typeinfo.BaseCharTypeInfo; @@ -49,15 +49,17 @@ + "Example:\n" + " > SELECT _FUNC_('abc', 'def') FROM src LIMIT 1;\n" + " 'abcdef'") -@VectorizedExpressions({StringConcatColCol.class, StringConcatColScalar.class, - StringConcatScalarCol.class}) +@VectorizedExpressions(preferredType="STRING...", + value={StringConcatColCol.class, StringConcatColScalar.class, StringConcatScalarCol.class}) public class GenericUDFConcat extends GenericUDF { - private transient ObjectInspector[] argumentOIs; + private transient PrimitiveObjectInspector[] argumentOIs; private transient StringConverter[] stringConverters; private transient PrimitiveCategory returnType = PrimitiveCategory.STRING; private transient BytesWritable[] bw; private transient GenericUDFUtils.StringHelper returnHelper; + private transient Object[] constants; + @Override public ObjectInspector initialize(ObjectInspector[] arguments) throws UDFArgumentException { @@ -68,16 +70,17 @@ public ObjectInspector initialize(ObjectInspector[] arguments) throws UDFArgumen // All CHAR/VARCHAR inputs: return VARCHAR // All BINARY inputs: return BINARY // Otherwise return STRING - argumentOIs = arguments; - PrimitiveCategory currentCategory; PrimitiveObjectInspector poi; boolean fixedLengthReturnValue = true; int returnLength = 0; // Only for char/varchar return types + + argumentOIs = new PrimitiveObjectInspector[arguments.length]; for (int idx = 0; idx < arguments.length; ++idx) { if (arguments[idx].getCategory() != Category.PRIMITIVE) { throw new UDFArgumentException("CONCAT only takes primitive arguments"); } + argumentOIs[idx] = (PrimitiveObjectInspector) arguments[idx]; poi = (PrimitiveObjectInspector)arguments[idx]; currentCategory = poi.getPrimitiveCategory(); if (idx == 0) { @@ -121,34 +124,59 @@ public ObjectInspector initialize(ObjectInspector[] arguments) throws UDFArgumen } } + constants = new Object[arguments.length]; + + boolean hasConstant = hasAnyConstants(arguments); + if (hasConstant) { + for (int i = 0; i < arguments.length; i++) { + if (arguments[i] instanceof ConstantObjectInspector) { + constants[i] = ((ConstantObjectInspector)arguments[i]).getWritableConstantValue(); + } + } + } + if (returnType == PrimitiveCategory.BINARY) { bw = new BytesWritable[arguments.length]; return PrimitiveObjectInspectorFactory.writableBinaryObjectInspector; - } else { - // treat all inputs as string, the return value will be converted to the appropriate type. - createStringConverters(); - returnHelper = new GenericUDFUtils.StringHelper(returnType); - BaseCharTypeInfo typeInfo; - switch (returnType) { - case STRING: - return PrimitiveObjectInspectorFactory.writableStringObjectInspector; - case CHAR: - typeInfo = TypeInfoFactory.getCharTypeInfo(returnLength); - return PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(typeInfo); - case VARCHAR: - typeInfo = TypeInfoFactory.getVarcharTypeInfo(returnLength); - return PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(typeInfo); - default: - throw new UDFArgumentException("Unexpected CONCAT return type of " + returnType); + } + // treat all inputs as string, the return value will be converted to the appropriate type. + stringConverters = createStringConverters(argumentOIs); + if (hasConstant) { + for (int i = 0; i < constants.length; i++) { + constants[i] = constants[i] == null ? null : stringConverters[i].convert(constants[i]); } } + returnHelper = new GenericUDFUtils.StringHelper(returnType); + BaseCharTypeInfo typeInfo; + switch (returnType) { + case STRING: + return PrimitiveObjectInspectorFactory.writableStringObjectInspector; + case CHAR: + typeInfo = TypeInfoFactory.getCharTypeInfo(returnLength); + return PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(typeInfo); + case VARCHAR: + typeInfo = TypeInfoFactory.getVarcharTypeInfo(returnLength); + return PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(typeInfo); + default: + throw new UDFArgumentException("Unexpected CONCAT return type of " + returnType); + } } - private void createStringConverters() { - stringConverters = new StringConverter[argumentOIs.length]; + private boolean hasAnyConstants(ObjectInspector[] argumentOIs) { + for (ObjectInspector argumentOI : argumentOIs) { + if (argumentOI instanceof ConstantObjectInspector) { + return true; + } + } + return false; + } + + private StringConverter[] createStringConverters(ObjectInspector[] argumentOIs) { + StringConverter[] stringConverters = new StringConverter[argumentOIs.length]; for (int idx = 0; idx < argumentOIs.length; ++idx) { stringConverters[idx] = new StringConverter((PrimitiveObjectInspector) argumentOIs[idx]); } + return stringConverters; } @Override @@ -160,11 +188,11 @@ public Object evaluate(DeferredObject[] arguments) throws HiveException { } } - public Object binaryEvaluate(DeferredObject[] arguments) throws HiveException { + private Object binaryEvaluate(DeferredObject[] arguments) throws HiveException { int len = 0; for (int idx = 0; idx < arguments.length; ++idx) { - bw[idx] = ((BinaryObjectInspector)argumentOIs[idx]) - .getPrimitiveWritableObject(arguments[idx].get()); + bw[idx] = (BytesWritable) (constants[idx] != null ? constants[idx] : + (argumentOIs[idx]).getPrimitiveWritableObject(arguments[idx].get())); if (bw[idx] == null){ return null; } @@ -181,19 +209,19 @@ public Object binaryEvaluate(DeferredObject[] arguments) throws HiveException { return new BytesWritable(out); } - public String stringEvaluate(DeferredObject[] arguments) throws HiveException { - StringBuilder sb = new StringBuilder(); + private transient StringBuilder builder = new StringBuilder(); + + private String stringEvaluate(DeferredObject[] arguments) throws HiveException { + builder.setLength(0); for (int idx = 0; idx < arguments.length; ++idx) { - String val = null; - if (arguments[idx] != null) { - val = (String) stringConverters[idx].convert(arguments[idx].get()); - } + String val = (String)(constants[idx] != null ? constants[idx] : + stringConverters[idx].convert(arguments[idx].get())); if (val == null) { return null; } - sb.append(val); + builder.append(val); } - return sb.toString(); + return builder.toString(); } @Override diff --git ql/src/test/org/apache/hadoop/hive/ql/exec/vector/TestVectorizationContext.java ql/src/test/org/apache/hadoop/hive/ql/exec/vector/TestVectorizationContext.java index 2329f52..42634d0 100644 --- ql/src/test/org/apache/hadoop/hive/ql/exec/vector/TestVectorizationContext.java +++ ql/src/test/org/apache/hadoop/hive/ql/exec/vector/TestVectorizationContext.java @@ -19,10 +19,12 @@ package org.apache.hadoop.hive.ql.exec.vector; import static org.junit.Assert.assertEquals; +import static org.junit.Assert.assertNull; import static org.junit.Assert.assertTrue; import java.sql.Timestamp; import java.util.ArrayList; +import java.util.Arrays; import java.util.HashMap; import java.util.List; import java.util.Map; @@ -1209,6 +1211,20 @@ public void testFoldConstantsForUnaryExpression() throws HiveException { assertTrue(constFoldNodeDesc instanceof ExprNodeConstantDesc); assertTrue(((HiveDecimal) (((ExprNodeConstantDesc)constFoldNodeDesc).getValue())).toString().equals("1")); - } + } + @Test + public void testPreferredTypes() { + assertEquals("[STRING, STRING, STRING]", Arrays.toString(VectorizationContext.toPreferredTypes(3, "STRING..."))); + assertEquals("[INT, STRING, STRING]", Arrays.toString(VectorizationContext.toPreferredTypes(3, "INT, STRING..."))); + assertEquals("[INT, STRING, INT]", Arrays.toString(VectorizationContext.toPreferredTypes(3, "INT, STRING..., INT"))); + + assertEquals("[STRING]", Arrays.toString(VectorizationContext.toPreferredTypes(1, "STRING..."))); + assertEquals("[INT]", Arrays.toString(VectorizationContext.toPreferredTypes(1, "INT, STRING..."))); + assertEquals("[INT, INT]", Arrays.toString(VectorizationContext.toPreferredTypes(2, "INT, STRING..., INT"))); + + assertNull(VectorizationContext.toPreferredTypes(2, "INT, INT, STRING..., INT")); + assertNull(VectorizationContext.toPreferredTypes(2, "INT, INT, INT")); + assertNull(VectorizationContext.toPreferredTypes(2, "INT")); + } } diff --git ql/src/test/queries/clientpositive/udf_concat_tez.q ql/src/test/queries/clientpositive/udf_concat_tez.q new file mode 100644 index 0000000..bad85de --- /dev/null +++ ql/src/test/queries/clientpositive/udf_concat_tez.q @@ -0,0 +1,8 @@ +set hive.vectorized.execution.enabled=true; + +create table lineitem (l_orderkey bigint, l_linestatus string) STORED AS ORC; +insert overwrite table lineitem select * from src tablesample (10 rows); + +explain +SELECT concat(l_orderkey, l_linestatus) FROM lineitem; +SELECT concat(l_orderkey, l_linestatus) FROM lineitem; diff --git ql/src/test/results/clientpositive/tez/udf_concat_tez.q.out ql/src/test/results/clientpositive/tez/udf_concat_tez.q.out new file mode 100644 index 0000000..4c0f8c5 --- /dev/null +++ ql/src/test/results/clientpositive/tez/udf_concat_tez.q.out @@ -0,0 +1,74 @@ +PREHOOK: query: create table lineitem (l_orderkey bigint, l_linestatus string) STORED AS ORC +PREHOOK: type: CREATETABLE +PREHOOK: Output: database:default +POSTHOOK: query: create table lineitem (l_orderkey bigint, l_linestatus string) STORED AS ORC +POSTHOOK: type: CREATETABLE +POSTHOOK: Output: database:default +POSTHOOK: Output: default@lineitem +PREHOOK: query: insert overwrite table lineitem select * from src tablesample (10 rows) +PREHOOK: type: QUERY +PREHOOK: Input: default@src +PREHOOK: Output: default@lineitem +POSTHOOK: query: insert overwrite table lineitem select * from src tablesample (10 rows) +POSTHOOK: type: QUERY +POSTHOOK: Input: default@src +POSTHOOK: Output: default@lineitem +POSTHOOK: Lineage: lineitem.l_linestatus SIMPLE [(src)src.FieldSchema(name:value, type:string, comment:default), ] +POSTHOOK: Lineage: lineitem.l_orderkey EXPRESSION [(src)src.FieldSchema(name:key, type:string, comment:default), ] +PREHOOK: query: explain +SELECT concat(l_orderkey, l_linestatus) FROM lineitem +PREHOOK: type: QUERY +POSTHOOK: query: explain +SELECT concat(l_orderkey, l_linestatus) FROM lineitem +POSTHOOK: type: QUERY +STAGE DEPENDENCIES: + Stage-1 is a root stage + Stage-0 is a root stage + +STAGE PLANS: + Stage: Stage-1 + Tez +#### A masked pattern was here #### + Vertices: + Map 1 + Map Operator Tree: + TableScan + alias: lineitem + Statistics: Num rows: 10 Data size: 980 Basic stats: COMPLETE Column stats: NONE + Select Operator + expressions: concat(l_orderkey, l_linestatus) (type: string) + outputColumnNames: _col0 + Statistics: Num rows: 10 Data size: 980 Basic stats: COMPLETE Column stats: NONE + File Output Operator + compressed: false + Statistics: Num rows: 10 Data size: 980 Basic stats: COMPLETE Column stats: NONE + table: + input format: org.apache.hadoop.mapred.TextInputFormat + output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat + serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe + Execution mode: vectorized + + Stage: Stage-0 + Fetch Operator + limit: -1 + Processor Tree: + ListSink + +PREHOOK: query: SELECT concat(l_orderkey, l_linestatus) FROM lineitem +PREHOOK: type: QUERY +PREHOOK: Input: default@lineitem +#### A masked pattern was here #### +POSTHOOK: query: SELECT concat(l_orderkey, l_linestatus) FROM lineitem +POSTHOOK: type: QUERY +POSTHOOK: Input: default@lineitem +#### A masked pattern was here #### +238val_238 +86val_86 +311val_311 +27val_27 +165val_165 +409val_409 +255val_255 +278val_278 +98val_98 +484val_484