ODIn/spark-instrumented-optimizer
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity
Apache Spark - A unified analytics engine for large-scale data processing
19821 commits 6 branches 181 tags 397 MiB
Scala 65.7%
Python 12.1%
Jupyter Notebook 7.3%
Java 7.3%
HiveQL 3.2%
Other 4%
Go to file
Michal Senkyr 0538f3b0ae [SPARK-18891][SQL] Support for Scala Map collection types 
## What changes were proposed in this pull request?

Add support for arbitrary Scala `Map` types in deserialization as well as a generic implicit encoder.

Used the builder approach as in #16541 to construct any provided `Map` type upon deserialization.

Please note that this PR also adds (ignored) tests for issue [SPARK-19104 CompileException with Map and Case Class in Spark 2.1.0](https://issues.apache.org/jira/browse/SPARK-19104) but doesn't solve it.

Added support for Java Maps in codegen code (encoders will be added in a different PR) with the following default implementations for interfaces/abstract classes:

* `java.util.Map`, `java.util.AbstractMap` => `java.util.HashMap`
* `java.util.SortedMap`, `java.util.NavigableMap` => `java.util.TreeMap`
* `java.util.concurrent.ConcurrentMap` => `java.util.concurrent.ConcurrentHashMap`
* `java.util.concurrent.ConcurrentNavigableMap` => `java.util.concurrent.ConcurrentSkipListMap`

Resulting codegen for `Seq(Map(1 -> 2)).toDS().map(identity).queryExecution.debug.codegen`:

```
/* 001 */ public Object generate(Object[] references) {
/* 002 */   return new GeneratedIterator(references);
/* 003 */ }
/* 004 */
/* 005 */ final class GeneratedIterator extends org.apache.spark.sql.execution.BufferedRowIterator {
/* 006 */   private Object[] references;
/* 007 */   private scala.collection.Iterator[] inputs;
/* 008 */   private scala.collection.Iterator inputadapter_input;
/* 009 */   private boolean CollectObjectsToMap_loopIsNull1;
/* 010 */   private int CollectObjectsToMap_loopValue0;
/* 011 */   private boolean CollectObjectsToMap_loopIsNull3;
/* 012 */   private int CollectObjectsToMap_loopValue2;
/* 013 */   private UnsafeRow deserializetoobject_result;
/* 014 */   private org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder deserializetoobject_holder;
/* 015 */   private org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter deserializetoobject_rowWriter;
/* 016 */   private scala.collection.immutable.Map mapelements_argValue;
/* 017 */   private UnsafeRow mapelements_result;
/* 018 */   private org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder mapelements_holder;
/* 019 */   private org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter mapelements_rowWriter;
/* 020 */   private UnsafeRow serializefromobject_result;
/* 021 */   private org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder serializefromobject_holder;
/* 022 */   private org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter serializefromobject_rowWriter;
/* 023 */   private org.apache.spark.sql.catalyst.expressions.codegen.UnsafeArrayWriter serializefromobject_arrayWriter;
/* 024 */   private org.apache.spark.sql.catalyst.expressions.codegen.UnsafeArrayWriter serializefromobject_arrayWriter1;
/* 025 */
/* 026 */   public GeneratedIterator(Object[] references) {
/* 027 */     this.references = references;
/* 028 */   }
/* 029 */
/* 030 */   public void init(int index, scala.collection.Iterator[] inputs) {
/* 031 */     partitionIndex = index;
/* 032 */     this.inputs = inputs;
/* 033 */     wholestagecodegen_init_0();
/* 034 */     wholestagecodegen_init_1();
/* 035 */
/* 036 */   }
/* 037 */
/* 038 */   private void wholestagecodegen_init_0() {
/* 039 */     inputadapter_input = inputs[0];
/* 040 */
/* 041 */     deserializetoobject_result = new UnsafeRow(1);
/* 042 */     this.deserializetoobject_holder = new org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder(deserializetoobject_result, 32);
/* 043 */     this.deserializetoobject_rowWriter = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter(deserializetoobject_holder, 1);
/* 044 */
/* 045 */     mapelements_result = new UnsafeRow(1);
/* 046 */     this.mapelements_holder = new org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder(mapelements_result, 32);
/* 047 */     this.mapelements_rowWriter = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter(mapelements_holder, 1);
/* 048 */     serializefromobject_result = new UnsafeRow(1);
/* 049 */     this.serializefromobject_holder = new org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder(serializefromobject_result, 32);
/* 050 */     this.serializefromobject_rowWriter = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter(serializefromobject_holder, 1);
/* 051 */     this.serializefromobject_arrayWriter = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeArrayWriter();
/* 052 */
/* 053 */   }
/* 054 */
/* 055 */   private void wholestagecodegen_init_1() {
/* 056 */     this.serializefromobject_arrayWriter1 = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeArrayWriter();
/* 057 */
/* 058 */   }
/* 059 */
/* 060 */   protected void processNext() throws java.io.IOException {
/* 061 */     while (inputadapter_input.hasNext() && !stopEarly()) {
/* 062 */       InternalRow inputadapter_row = (InternalRow) inputadapter_input.next();
/* 063 */       boolean inputadapter_isNull = inputadapter_row.isNullAt(0);
/* 064 */       MapData inputadapter_value = inputadapter_isNull ? null : (inputadapter_row.getMap(0));
/* 065 */
/* 066 */       boolean deserializetoobject_isNull1 = true;
/* 067 */       ArrayData deserializetoobject_value1 = null;
/* 068 */       if (!inputadapter_isNull) {
/* 069 */         deserializetoobject_isNull1 = false;
/* 070 */         if (!deserializetoobject_isNull1) {
/* 071 */           Object deserializetoobject_funcResult = null;
/* 072 */           deserializetoobject_funcResult = inputadapter_value.keyArray();
/* 073 */           if (deserializetoobject_funcResult == null) {
/* 074 */             deserializetoobject_isNull1 = true;
/* 075 */           } else {
/* 076 */             deserializetoobject_value1 = (ArrayData) deserializetoobject_funcResult;
/* 077 */           }
/* 078 */
/* 079 */         }
/* 080 */         deserializetoobject_isNull1 = deserializetoobject_value1 == null;
/* 081 */       }
/* 082 */
/* 083 */       boolean deserializetoobject_isNull3 = true;
/* 084 */       ArrayData deserializetoobject_value3 = null;
/* 085 */       if (!inputadapter_isNull) {
/* 086 */         deserializetoobject_isNull3 = false;
/* 087 */         if (!deserializetoobject_isNull3) {
/* 088 */           Object deserializetoobject_funcResult1 = null;
/* 089 */           deserializetoobject_funcResult1 = inputadapter_value.valueArray();
/* 090 */           if (deserializetoobject_funcResult1 == null) {
/* 091 */             deserializetoobject_isNull3 = true;
/* 092 */           } else {
/* 093 */             deserializetoobject_value3 = (ArrayData) deserializetoobject_funcResult1;
/* 094 */           }
/* 095 */
/* 096 */         }
/* 097 */         deserializetoobject_isNull3 = deserializetoobject_value3 == null;
/* 098 */       }
/* 099 */       scala.collection.immutable.Map deserializetoobject_value = null;
/* 100 */
/* 101 */       if ((deserializetoobject_isNull1 && !deserializetoobject_isNull3) ||
/* 102 */         (!deserializetoobject_isNull1 && deserializetoobject_isNull3)) {
/* 103 */         throw new RuntimeException("Invalid state: Inconsistent nullability of key-value");
/* 104 */       }
/* 105 */
/* 106 */       if (!deserializetoobject_isNull1) {
/* 107 */         if (deserializetoobject_value1.numElements() != deserializetoobject_value3.numElements()) {
/* 108 */           throw new RuntimeException("Invalid state: Inconsistent lengths of key-value arrays");
/* 109 */         }
/* 110 */         int deserializetoobject_dataLength = deserializetoobject_value1.numElements();
/* 111 */
/* 112 */         scala.collection.mutable.Builder CollectObjectsToMap_builderValue5 = scala.collection.immutable.Map$.MODULE$.newBuilder();
/* 113 */         CollectObjectsToMap_builderValue5.sizeHint(deserializetoobject_dataLength);
/* 114 */
/* 115 */         int deserializetoobject_loopIndex = 0;
/* 116 */         while (deserializetoobject_loopIndex < deserializetoobject_dataLength) {
/* 117 */           CollectObjectsToMap_loopValue0 = (int) (deserializetoobject_value1.getInt(deserializetoobject_loopIndex));
/* 118 */           CollectObjectsToMap_loopValue2 = (int) (deserializetoobject_value3.getInt(deserializetoobject_loopIndex));
/* 119 */           CollectObjectsToMap_loopIsNull1 = deserializetoobject_value1.isNullAt(deserializetoobject_loopIndex);
/* 120 */           CollectObjectsToMap_loopIsNull3 = deserializetoobject_value3.isNullAt(deserializetoobject_loopIndex);
/* 121 */
/* 122 */           if (CollectObjectsToMap_loopIsNull1) {
/* 123 */             throw new RuntimeException("Found null in map key!");
/* 124 */           }
/* 125 */
/* 126 */           scala.Tuple2 CollectObjectsToMap_loopValue4;
/* 127 */
/* 128 */           if (CollectObjectsToMap_loopIsNull3) {
/* 129 */             CollectObjectsToMap_loopValue4 = new scala.Tuple2(CollectObjectsToMap_loopValue0, null);
/* 130 */           } else {
/* 131 */             CollectObjectsToMap_loopValue4 = new scala.Tuple2(CollectObjectsToMap_loopValue0, CollectObjectsToMap_loopValue2);
/* 132 */           }
/* 133 */
/* 134 */           CollectObjectsToMap_builderValue5.$plus$eq(CollectObjectsToMap_loopValue4);
/* 135 */
/* 136 */           deserializetoobject_loopIndex += 1;
/* 137 */         }
/* 138 */
/* 139 */         deserializetoobject_value = (scala.collection.immutable.Map) CollectObjectsToMap_builderValue5.result();
/* 140 */       }
/* 141 */
/* 142 */       boolean mapelements_isNull = true;
/* 143 */       scala.collection.immutable.Map mapelements_value = null;
/* 144 */       if (!false) {
/* 145 */         mapelements_argValue = deserializetoobject_value;
/* 146 */
/* 147 */         mapelements_isNull = false;
/* 148 */         if (!mapelements_isNull) {
/* 149 */           Object mapelements_funcResult = null;
/* 150 */           mapelements_funcResult = ((scala.Function1) references[0]).apply(mapelements_argValue);
/* 151 */           if (mapelements_funcResult == null) {
/* 152 */             mapelements_isNull = true;
/* 153 */           } else {
/* 154 */             mapelements_value = (scala.collection.immutable.Map) mapelements_funcResult;
/* 155 */           }
/* 156 */
/* 157 */         }
/* 158 */         mapelements_isNull = mapelements_value == null;
/* 159 */       }
/* 160 */
/* 161 */       MapData serializefromobject_value = null;
/* 162 */       if (!mapelements_isNull) {
/* 163 */         final int serializefromobject_length = mapelements_value.size();
/* 164 */         final Object[] serializefromobject_convertedKeys = new Object[serializefromobject_length];
/* 165 */         final Object[] serializefromobject_convertedValues = new Object[serializefromobject_length];
/* 166 */         int serializefromobject_index = 0;
/* 167 */         final scala.collection.Iterator serializefromobject_entries = mapelements_value.iterator();
/* 168 */         while(serializefromobject_entries.hasNext()) {
/* 169 */           final scala.Tuple2 serializefromobject_entry = (scala.Tuple2) serializefromobject_entries.next();
/* 170 */           int ExternalMapToCatalyst_key1 = (Integer) serializefromobject_entry._1();
/* 171 */           int ExternalMapToCatalyst_value1 = (Integer) serializefromobject_entry._2();
/* 172 */
/* 173 */           boolean ExternalMapToCatalyst_value_isNull1 = false;
/* 174 */
/* 175 */           if (false) {
/* 176 */             throw new RuntimeException("Cannot use null as map key!");
/* 177 */           } else {
/* 178 */             serializefromobject_convertedKeys[serializefromobject_index] = (Integer) ExternalMapToCatalyst_key1;
/* 179 */           }
/* 180 */
/* 181 */           if (false) {
/* 182 */             serializefromobject_convertedValues[serializefromobject_index] = null;
/* 183 */           } else {
/* 184 */             serializefromobject_convertedValues[serializefromobject_index] = (Integer) ExternalMapToCatalyst_value1;
/* 185 */           }
/* 186 */
/* 187 */           serializefromobject_index++;
/* 188 */         }
/* 189 */
/* 190 */         serializefromobject_value = new org.apache.spark.sql.catalyst.util.ArrayBasedMapData(new org.apache.spark.sql.catalyst.util.GenericArrayData(serializefromobject_convertedKeys), new org.apache.spark.sql.catalyst.util.GenericArrayData(serializefromobject_convertedValues));
/* 191 */       }
/* 192 */       serializefromobject_holder.reset();
/* 193 */
/* 194 */       serializefromobject_rowWriter.zeroOutNullBytes();
/* 195 */
/* 196 */       if (mapelements_isNull) {
/* 197 */         serializefromobject_rowWriter.setNullAt(0);
/* 198 */       } else {
/* 199 */         // Remember the current cursor so that we can calculate how many bytes are
/* 200 */         // written later.
/* 201 */         final int serializefromobject_tmpCursor = serializefromobject_holder.cursor;
/* 202 */
/* 203 */         if (serializefromobject_value instanceof UnsafeMapData) {
/* 204 */           final int serializefromobject_sizeInBytes = ((UnsafeMapData) serializefromobject_value).getSizeInBytes();
/* 205 */           // grow the global buffer before writing data.
/* 206 */           serializefromobject_holder.grow(serializefromobject_sizeInBytes);
/* 207 */           ((UnsafeMapData) serializefromobject_value).writeToMemory(serializefromobject_holder.buffer, serializefromobject_holder.cursor);
/* 208 */           serializefromobject_holder.cursor += serializefromobject_sizeInBytes;
/* 209 */
/* 210 */         } else {
/* 211 */           final ArrayData serializefromobject_keys = serializefromobject_value.keyArray();
/* 212 */           final ArrayData serializefromobject_values = serializefromobject_value.valueArray();
/* 213 */
/* 214 */           // preserve 8 bytes to write the key array numBytes later.
/* 215 */           serializefromobject_holder.grow(8);
/* 216 */           serializefromobject_holder.cursor += 8;
/* 217 */
/* 218 */           // Remember the current cursor so that we can write numBytes of key array later.
/* 219 */           final int serializefromobject_tmpCursor1 = serializefromobject_holder.cursor;
/* 220 */
/* 221 */           if (serializefromobject_keys instanceof UnsafeArrayData) {
/* 222 */             final int serializefromobject_sizeInBytes1 = ((UnsafeArrayData) serializefromobject_keys).getSizeInBytes();
/* 223 */             // grow the global buffer before writing data.
/* 224 */             serializefromobject_holder.grow(serializefromobject_sizeInBytes1);
/* 225 */             ((UnsafeArrayData) serializefromobject_keys).writeToMemory(serializefromobject_holder.buffer, serializefromobject_holder.cursor);
/* 226 */             serializefromobject_holder.cursor += serializefromobject_sizeInBytes1;
/* 227 */
/* 228 */           } else {
/* 229 */             final int serializefromobject_numElements = serializefromobject_keys.numElements();
/* 230 */             serializefromobject_arrayWriter.initialize(serializefromobject_holder, serializefromobject_numElements, 4);
/* 231 */
/* 232 */             for (int serializefromobject_index1 = 0; serializefromobject_index1 < serializefromobject_numElements; serializefromobject_index1++) {
/* 233 */               if (serializefromobject_keys.isNullAt(serializefromobject_index1)) {
/* 234 */                 serializefromobject_arrayWriter.setNullInt(serializefromobject_index1);
/* 235 */               } else {
/* 236 */                 final int serializefromobject_element = serializefromobject_keys.getInt(serializefromobject_index1);
/* 237 */                 serializefromobject_arrayWriter.write(serializefromobject_index1, serializefromobject_element);
/* 238 */               }
/* 239 */             }
/* 240 */           }
/* 241 */
/* 242 */           // Write the numBytes of key array into the first 8 bytes.
/* 243 */           Platform.putLong(serializefromobject_holder.buffer, serializefromobject_tmpCursor1 - 8, serializefromobject_holder.cursor - serializefromobject_tmpCursor1);
/* 244 */
/* 245 */           if (serializefromobject_values instanceof UnsafeArrayData) {
/* 246 */             final int serializefromobject_sizeInBytes2 = ((UnsafeArrayData) serializefromobject_values).getSizeInBytes();
/* 247 */             // grow the global buffer before writing data.
/* 248 */             serializefromobject_holder.grow(serializefromobject_sizeInBytes2);
/* 249 */             ((UnsafeArrayData) serializefromobject_values).writeToMemory(serializefromobject_holder.buffer, serializefromobject_holder.cursor);
/* 250 */             serializefromobject_holder.cursor += serializefromobject_sizeInBytes2;
/* 251 */
/* 252 */           } else {
/* 253 */             final int serializefromobject_numElements1 = serializefromobject_values.numElements();
/* 254 */             serializefromobject_arrayWriter1.initialize(serializefromobject_holder, serializefromobject_numElements1, 4);
/* 255 */
/* 256 */             for (int serializefromobject_index2 = 0; serializefromobject_index2 < serializefromobject_numElements1; serializefromobject_index2++) {
/* 257 */               if (serializefromobject_values.isNullAt(serializefromobject_index2)) {
/* 258 */                 serializefromobject_arrayWriter1.setNullInt(serializefromobject_index2);
/* 259 */               } else {
/* 260 */                 final int serializefromobject_element1 = serializefromobject_values.getInt(serializefromobject_index2);
/* 261 */                 serializefromobject_arrayWriter1.write(serializefromobject_index2, serializefromobject_element1);
/* 262 */               }
/* 263 */             }
/* 264 */           }
/* 265 */
/* 266 */         }
/* 267 */
/* 268 */         serializefromobject_rowWriter.setOffsetAndSize(0, serializefromobject_tmpCursor, serializefromobject_holder.cursor - serializefromobject_tmpCursor);
/* 269 */       }
/* 270 */       serializefromobject_result.setTotalSize(serializefromobject_holder.totalSize());
/* 271 */       append(serializefromobject_result);
/* 272 */       if (shouldStop()) return;
/* 273 */     }
/* 274 */   }
/* 275 */ }
```

Codegen for `java.util.Map`:

```
/* 001 */ public Object generate(Object[] references) {
/* 002 */   return new GeneratedIterator(references);
/* 003 */ }
/* 004 */
/* 005 */ final class GeneratedIterator extends org.apache.spark.sql.execution.BufferedRowIterator {
/* 006 */   private Object[] references;
/* 007 */   private scala.collection.Iterator[] inputs;
/* 008 */   private scala.collection.Iterator inputadapter_input;
/* 009 */   private boolean CollectObjectsToMap_loopIsNull1;
/* 010 */   private int CollectObjectsToMap_loopValue0;
/* 011 */   private boolean CollectObjectsToMap_loopIsNull3;
/* 012 */   private int CollectObjectsToMap_loopValue2;
/* 013 */   private UnsafeRow deserializetoobject_result;
/* 014 */   private org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder deserializetoobject_holder;
/* 015 */   private org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter deserializetoobject_rowWriter;
/* 016 */   private java.util.HashMap mapelements_argValue;
/* 017 */   private UnsafeRow mapelements_result;
/* 018 */   private org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder mapelements_holder;
/* 019 */   private org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter mapelements_rowWriter;
/* 020 */   private UnsafeRow serializefromobject_result;
/* 021 */   private org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder serializefromobject_holder;
/* 022 */   private org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter serializefromobject_rowWriter;
/* 023 */   private org.apache.spark.sql.catalyst.expressions.codegen.UnsafeArrayWriter serializefromobject_arrayWriter;
/* 024 */   private org.apache.spark.sql.catalyst.expressions.codegen.UnsafeArrayWriter serializefromobject_arrayWriter1;
/* 025 */
/* 026 */   public GeneratedIterator(Object[] references) {
/* 027 */     this.references = references;
/* 028 */   }
/* 029 */
/* 030 */   public void init(int index, scala.collection.Iterator[] inputs) {
/* 031 */     partitionIndex = index;
/* 032 */     this.inputs = inputs;
/* 033 */     wholestagecodegen_init_0();
/* 034 */     wholestagecodegen_init_1();
/* 035 */
/* 036 */   }
/* 037 */
/* 038 */   private void wholestagecodegen_init_0() {
/* 039 */     inputadapter_input = inputs[0];
/* 040 */
/* 041 */     deserializetoobject_result = new UnsafeRow(1);
/* 042 */     this.deserializetoobject_holder = new org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder(deserializetoobject_result, 32);
/* 043 */     this.deserializetoobject_rowWriter = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter(deserializetoobject_holder, 1);
/* 044 */
/* 045 */     mapelements_result = new UnsafeRow(1);
/* 046 */     this.mapelements_holder = new org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder(mapelements_result, 32);
/* 047 */     this.mapelements_rowWriter = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter(mapelements_holder, 1);
/* 048 */     serializefromobject_result = new UnsafeRow(1);
/* 049 */     this.serializefromobject_holder = new org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder(serializefromobject_result, 32);
/* 050 */     this.serializefromobject_rowWriter = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter(serializefromobject_holder, 1);
/* 051 */     this.serializefromobject_arrayWriter = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeArrayWriter();
/* 052 */
/* 053 */   }
/* 054 */
/* 055 */   private void wholestagecodegen_init_1() {
/* 056 */     this.serializefromobject_arrayWriter1 = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeArrayWriter();
/* 057 */
/* 058 */   }
/* 059 */
/* 060 */   protected void processNext() throws java.io.IOException {
/* 061 */     while (inputadapter_input.hasNext() && !stopEarly()) {
/* 062 */       InternalRow inputadapter_row = (InternalRow) inputadapter_input.next();
/* 063 */       boolean inputadapter_isNull = inputadapter_row.isNullAt(0);
/* 064 */       MapData inputadapter_value = inputadapter_isNull ? null : (inputadapter_row.getMap(0));
/* 065 */
/* 066 */       boolean deserializetoobject_isNull1 = true;
/* 067 */       ArrayData deserializetoobject_value1 = null;
/* 068 */       if (!inputadapter_isNull) {
/* 069 */         deserializetoobject_isNull1 = false;
/* 070 */         if (!deserializetoobject_isNull1) {
/* 071 */           Object deserializetoobject_funcResult = null;
/* 072 */           deserializetoobject_funcResult = inputadapter_value.keyArray();
/* 073 */           if (deserializetoobject_funcResult == null) {
/* 074 */             deserializetoobject_isNull1 = true;
/* 075 */           } else {
/* 076 */             deserializetoobject_value1 = (ArrayData) deserializetoobject_funcResult;
/* 077 */           }
/* 078 */
/* 079 */         }
/* 080 */         deserializetoobject_isNull1 = deserializetoobject_value1 == null;
/* 081 */       }
/* 082 */
/* 083 */       boolean deserializetoobject_isNull3 = true;
/* 084 */       ArrayData deserializetoobject_value3 = null;
/* 085 */       if (!inputadapter_isNull) {
/* 086 */         deserializetoobject_isNull3 = false;
/* 087 */         if (!deserializetoobject_isNull3) {
/* 088 */           Object deserializetoobject_funcResult1 = null;
/* 089 */           deserializetoobject_funcResult1 = inputadapter_value.valueArray();
/* 090 */           if (deserializetoobject_funcResult1 == null) {
/* 091 */             deserializetoobject_isNull3 = true;
/* 092 */           } else {
/* 093 */             deserializetoobject_value3 = (ArrayData) deserializetoobject_funcResult1;
/* 094 */           }
/* 095 */
/* 096 */         }
/* 097 */         deserializetoobject_isNull3 = deserializetoobject_value3 == null;
/* 098 */       }
/* 099 */       java.util.HashMap deserializetoobject_value = null;
/* 100 */
/* 101 */       if ((deserializetoobject_isNull1 && !deserializetoobject_isNull3) ||
/* 102 */         (!deserializetoobject_isNull1 && deserializetoobject_isNull3)) {
/* 103 */         throw new RuntimeException("Invalid state: Inconsistent nullability of key-value");
/* 104 */       }
/* 105 */
/* 106 */       if (!deserializetoobject_isNull1) {
/* 107 */         if (deserializetoobject_value1.numElements() != deserializetoobject_value3.numElements()) {
/* 108 */           throw new RuntimeException("Invalid state: Inconsistent lengths of key-value arrays");
/* 109 */         }
/* 110 */         int deserializetoobject_dataLength = deserializetoobject_value1.numElements();
/* 111 */         java.util.Map CollectObjectsToMap_builderValue5 = new java.util.HashMap(deserializetoobject_dataLength);
/* 112 */
/* 113 */         int deserializetoobject_loopIndex = 0;
/* 114 */         while (deserializetoobject_loopIndex < deserializetoobject_dataLength) {
/* 115 */           CollectObjectsToMap_loopValue0 = (int) (deserializetoobject_value1.getInt(deserializetoobject_loopIndex));
/* 116 */           CollectObjectsToMap_loopValue2 = (int) (deserializetoobject_value3.getInt(deserializetoobject_loopIndex));
/* 117 */           CollectObjectsToMap_loopIsNull1 = deserializetoobject_value1.isNullAt(deserializetoobject_loopIndex);
/* 118 */           CollectObjectsToMap_loopIsNull3 = deserializetoobject_value3.isNullAt(deserializetoobject_loopIndex);
/* 119 */
/* 120 */           if (CollectObjectsToMap_loopIsNull1) {
/* 121 */             throw new RuntimeException("Found null in map key!");
/* 122 */           }
/* 123 */
/* 124 */           CollectObjectsToMap_builderValue5.put(CollectObjectsToMap_loopValue0, CollectObjectsToMap_loopValue2);
/* 125 */
/* 126 */           deserializetoobject_loopIndex += 1;
/* 127 */         }
/* 128 */
/* 129 */         deserializetoobject_value = (java.util.HashMap) CollectObjectsToMap_builderValue5;
/* 130 */       }
/* 131 */
/* 132 */       boolean mapelements_isNull = true;
/* 133 */       java.util.HashMap mapelements_value = null;
/* 134 */       if (!false) {
/* 135 */         mapelements_argValue = deserializetoobject_value;
/* 136 */
/* 137 */         mapelements_isNull = false;
/* 138 */         if (!mapelements_isNull) {
/* 139 */           Object mapelements_funcResult = null;
/* 140 */           mapelements_funcResult = ((scala.Function1) references[0]).apply(mapelements_argValue);
/* 141 */           if (mapelements_funcResult == null) {
/* 142 */             mapelements_isNull = true;
/* 143 */           } else {
/* 144 */             mapelements_value = (java.util.HashMap) mapelements_funcResult;
/* 145 */           }
/* 146 */
/* 147 */         }
/* 148 */         mapelements_isNull = mapelements_value == null;
/* 149 */       }
/* 150 */
/* 151 */       MapData serializefromobject_value = null;
/* 152 */       if (!mapelements_isNull) {
/* 153 */         final int serializefromobject_length = mapelements_value.size();
/* 154 */         final Object[] serializefromobject_convertedKeys = new Object[serializefromobject_length];
/* 155 */         final Object[] serializefromobject_convertedValues = new Object[serializefromobject_length];
/* 156 */         int serializefromobject_index = 0;
/* 157 */         final java.util.Iterator serializefromobject_entries = mapelements_value.entrySet().iterator();
/* 158 */         while(serializefromobject_entries.hasNext()) {
/* 159 */           final java.util.Map$Entry serializefromobject_entry = (java.util.Map$Entry) serializefromobject_entries.next();
/* 160 */           int ExternalMapToCatalyst_key1 = (Integer) serializefromobject_entry.getKey();
/* 161 */           int ExternalMapToCatalyst_value1 = (Integer) serializefromobject_entry.getValue();
/* 162 */
/* 163 */           boolean ExternalMapToCatalyst_value_isNull1 = false;
/* 164 */
/* 165 */           if (false) {
/* 166 */             throw new RuntimeException("Cannot use null as map key!");
/* 167 */           } else {
/* 168 */             serializefromobject_convertedKeys[serializefromobject_index] = (Integer) ExternalMapToCatalyst_key1;
/* 169 */           }
/* 170 */
/* 171 */           if (false) {
/* 172 */             serializefromobject_convertedValues[serializefromobject_index] = null;
/* 173 */           } else {
/* 174 */             serializefromobject_convertedValues[serializefromobject_index] = (Integer) ExternalMapToCatalyst_value1;
/* 175 */           }
/* 176 */
/* 177 */           serializefromobject_index++;
/* 178 */         }
/* 179 */
/* 180 */         serializefromobject_value = new org.apache.spark.sql.catalyst.util.ArrayBasedMapData(new org.apache.spark.sql.catalyst.util.GenericArrayData(serializefromobject_convertedKeys), new org.apache.spark.sql.catalyst.util.GenericArrayData(serializefromobject_convertedValues));
/* 181 */       }
/* 182 */       serializefromobject_holder.reset();
/* 183 */
/* 184 */       serializefromobject_rowWriter.zeroOutNullBytes();
/* 185 */
/* 186 */       if (mapelements_isNull) {
/* 187 */         serializefromobject_rowWriter.setNullAt(0);
/* 188 */       } else {
/* 189 */         // Remember the current cursor so that we can calculate how many bytes are
/* 190 */         // written later.
/* 191 */         final int serializefromobject_tmpCursor = serializefromobject_holder.cursor;
/* 192 */
/* 193 */         if (serializefromobject_value instanceof UnsafeMapData) {
/* 194 */           final int serializefromobject_sizeInBytes = ((UnsafeMapData) serializefromobject_value).getSizeInBytes();
/* 195 */           // grow the global buffer before writing data.
/* 196 */           serializefromobject_holder.grow(serializefromobject_sizeInBytes);
/* 197 */           ((UnsafeMapData) serializefromobject_value).writeToMemory(serializefromobject_holder.buffer, serializefromobject_holder.cursor);
/* 198 */           serializefromobject_holder.cursor += serializefromobject_sizeInBytes;
/* 199 */
/* 200 */         } else {
/* 201 */           final ArrayData serializefromobject_keys = serializefromobject_value.keyArray();
/* 202 */           final ArrayData serializefromobject_values = serializefromobject_value.valueArray();
/* 203 */
/* 204 */           // preserve 8 bytes to write the key array numBytes later.
/* 205 */           serializefromobject_holder.grow(8);
/* 206 */           serializefromobject_holder.cursor += 8;
/* 207 */
/* 208 */           // Remember the current cursor so that we can write numBytes of key array later.
/* 209 */           final int serializefromobject_tmpCursor1 = serializefromobject_holder.cursor;
/* 210 */
/* 211 */           if (serializefromobject_keys instanceof UnsafeArrayData) {
/* 212 */             final int serializefromobject_sizeInBytes1 = ((UnsafeArrayData) serializefromobject_keys).getSizeInBytes();
/* 213 */             // grow the global buffer before writing data.
/* 214 */             serializefromobject_holder.grow(serializefromobject_sizeInBytes1);
/* 215 */             ((UnsafeArrayData) serializefromobject_keys).writeToMemory(serializefromobject_holder.buffer, serializefromobject_holder.cursor);
/* 216 */             serializefromobject_holder.cursor += serializefromobject_sizeInBytes1;
/* 217 */
/* 218 */           } else {
/* 219 */             final int serializefromobject_numElements = serializefromobject_keys.numElements();
/* 220 */             serializefromobject_arrayWriter.initialize(serializefromobject_holder, serializefromobject_numElements, 4);
/* 221 */
/* 222 */             for (int serializefromobject_index1 = 0; serializefromobject_index1 < serializefromobject_numElements; serializefromobject_index1++) {
/* 223 */               if (serializefromobject_keys.isNullAt(serializefromobject_index1)) {
/* 224 */                 serializefromobject_arrayWriter.setNullInt(serializefromobject_index1);
/* 225 */               } else {
/* 226 */                 final int serializefromobject_element = serializefromobject_keys.getInt(serializefromobject_index1);
/* 227 */                 serializefromobject_arrayWriter.write(serializefromobject_index1, serializefromobject_element);
/* 228 */               }
/* 229 */             }
/* 230 */           }
/* 231 */
/* 232 */           // Write the numBytes of key array into the first 8 bytes.
/* 233 */           Platform.putLong(serializefromobject_holder.buffer, serializefromobject_tmpCursor1 - 8, serializefromobject_holder.cursor - serializefromobject_tmpCursor1);
/* 234 */
/* 235 */           if (serializefromobject_values instanceof UnsafeArrayData) {
/* 236 */             final int serializefromobject_sizeInBytes2 = ((UnsafeArrayData) serializefromobject_values).getSizeInBytes();
/* 237 */             // grow the global buffer before writing data.
/* 238 */             serializefromobject_holder.grow(serializefromobject_sizeInBytes2);
/* 239 */             ((UnsafeArrayData) serializefromobject_values).writeToMemory(serializefromobject_holder.buffer, serializefromobject_holder.cursor);
/* 240 */             serializefromobject_holder.cursor += serializefromobject_sizeInBytes2;
/* 241 */
/* 242 */           } else {
/* 243 */             final int serializefromobject_numElements1 = serializefromobject_values.numElements();
/* 244 */             serializefromobject_arrayWriter1.initialize(serializefromobject_holder, serializefromobject_numElements1, 4);
/* 245 */
/* 246 */             for (int serializefromobject_index2 = 0; serializefromobject_index2 < serializefromobject_numElements1; serializefromobject_index2++) {
/* 247 */               if (serializefromobject_values.isNullAt(serializefromobject_index2)) {
/* 248 */                 serializefromobject_arrayWriter1.setNullInt(serializefromobject_index2);
/* 249 */               } else {
/* 250 */                 final int serializefromobject_element1 = serializefromobject_values.getInt(serializefromobject_index2);
/* 251 */                 serializefromobject_arrayWriter1.write(serializefromobject_index2, serializefromobject_element1);
/* 252 */               }
/* 253 */             }
/* 254 */           }
/* 255 */
/* 256 */         }
/* 257 */
/* 258 */         serializefromobject_rowWriter.setOffsetAndSize(0, serializefromobject_tmpCursor, serializefromobject_holder.cursor - serializefromobject_tmpCursor);
/* 259 */       }
/* 260 */       serializefromobject_result.setTotalSize(serializefromobject_holder.totalSize());
/* 261 */       append(serializefromobject_result);
/* 262 */       if (shouldStop()) return;
/* 263 */     }
/* 264 */   }
/* 265 */ }
```

## How was this patch tested?

```
build/mvn -DskipTests clean package && dev/run-tests
```

Additionally in Spark shell:

```
scala> Seq(collection.mutable.HashMap(1 -> 2, 2 -> 3)).toDS().map(_ += (3 -> 4)).collect()
res0: Array[scala.collection.mutable.HashMap[Int,Int]] = Array(Map(2 -> 3, 1 -> 2, 3 -> 4))
```

Author: Michal Senkyr <mike.senkyr@gmail.com>
Author: Michal Šenkýř <mike.senkyr@gmail.com>

Closes #16986 from michalsenkyr/dataset-map-builder.
2017-06-12 08:47:01 +08:00
.github [SPARK-18073][DOCS][WIP] Migrate wiki to spark.apache.org web site 2016-11-23 11:25:47 +00:00
assembly [SPARK-7481][BUILD] Add spark-hadoop-cloud module to pull in object store access. 2017-05-07 10:15:31 +01:00
bin [SPARK-20613] Remove excess quotes in Windows executable 2017-05-05 08:30:42 -07:00
build [SPARK-19550][BUILD][CORE][WIP] Remove Java 7 support 2017-02-16 12:32:45 +00:00
common [SPARK-20641][CORE] Add key-value store abstraction and LevelDB implementation. 2017-06-06 13:39:10 -05:00
conf [SPARK-20995][CORE] Spark-env.sh.template' should add 'YARN_CONF_DIR' configuration instructions. 2017-06-09 09:26:30 +01:00
core [SPARK-20997][CORE] driver-cores' standalone or Mesos or YARN in Cluster deploy mode only. 2017-06-09 14:26:54 +01:00
data [SPARK-16421][EXAMPLES][ML] Improve ML Example Outputs 2016-08-05 20:57:46 +01:00
dev [SPARK-13933][BUILD] Update hadoop-2.7 profile's curator version to 2.7.1 2017-06-11 10:05:47 +01:00
docs [SPARK-21000][MESOS] Add Mesos labels support to the Spark Dispatcher 2017-06-11 09:49:39 +01:00
examples Fix bug in JavaRegressionMetricsExample. 2017-06-09 10:49:04 +01:00
external [SPARK-19185][DSTREAM] Make Kafka consumer cache configurable 2017-06-08 09:55:43 -07:00
graphx [SPARK-20523][BUILD] Clean up build warnings for 2.2.0 release 2017-05-03 10:18:35 +01:00
hadoop-cloud [SPARK-7481][BUILD] Add spark-hadoop-cloud module to pull in object store access. 2017-05-07 10:15:31 +01:00
launcher [SPARK-20922][CORE] Add whitelist of classes that can be deserialized by the launcher. 2017-06-01 14:44:34 -07:00
licenses [MINOR][BUILD] Add modernizr MIT license; specify "2014 and onwards" in license copyright 2016-06-04 21:41:27 +01:00
mllib [SPARK-19762][ML] Hierarchy for consolidating ML aggregator/loss code 2017-06-05 10:32:17 +01:00
mllib-local [SPARK-20677][MLLIB][ML] Follow-up to ALS recommend-all performance PRs 2017-05-16 10:59:34 +02:00
project [SPARK-20641][CORE] Add key-value store abstraction and LevelDB implementation. 2017-06-06 13:39:10 -05:00
python [SPARK-21042][SQL] Document Dataset.union is resolution by position 2017-06-09 18:29:33 -07:00
R [SPARK-20877][SPARKR][FOLLOWUP] clean up after test move 2017-06-11 03:00:44 -07:00
repl [SPARK-20548][FLAKY-TEST] share one REPL instance among REPL test cases 2017-05-10 00:09:35 +08:00
resource-managers [SPARK-21000][MESOS] Add Mesos labels support to the Spark Dispatcher 2017-06-11 09:49:39 +01:00
sbin [SPARK-19083] sbin/start-history-server.sh script use of $@ without quotes 2017-01-06 09:57:49 -08:00
sql [SPARK-18891][SQL] Support for Scala Map collection types 2017-06-12 08:47:01 +08:00
streaming [SPARK-20935][STREAMING] Always close WriteAheadLog and make it idempotent 2017-06-11 09:54:57 +01:00
tools [SPARK-20453] Bump master branch version to 2.3.0-SNAPSHOT 2017-04-24 21:48:04 -07:00
.gitattributes [SPARK-3870] EOL character enforcement 2014-10-31 12:39:52 -07:00
.gitignore [SPARK-19562][BUILD] Added exclude for dev/pr-deps to gitignore 2017-02-13 11:22:31 +00:00
.travis.yml [SPARK-19801][BUILD] Remove JDK7 from Travis CI 2017-03-03 12:00:54 +01:00
appveyor.yml [SPARK-20543][SPARKR][FOLLOWUP] Don't skip tests on AppVeyor 2017-05-07 13:10:10 -07:00
CONTRIBUTING.md [SPARK-18073][DOCS][WIP] Migrate wiki to spark.apache.org web site 2016-11-23 11:25:47 +00:00
LICENSE [SPARK-20759] SCALA_VERSION in _config.yml should be consistent with pom.xml 2017-05-19 15:26:39 +01:00
NOTICE [SPARK-18262][BUILD][SQL] JSON.org license is now CatX 2016-11-10 10:20:03 -08:00
pom.xml [SPARK-13933][BUILD] Update hadoop-2.7 profile's curator version to 2.7.1 2017-06-11 10:05:47 +01:00
README.md [MINOR][DOCS] Replace non-breaking space to normal spaces that breaks rendering markdown 2017-04-03 10:09:11 +01:00
scalastyle-config.xml [SPARK-13747][CORE] Add ThreadUtils.awaitReady and disallow Await.ready 2017-05-17 17:21:46 -07:00

README.md

Apache Spark

Spark is a fast and general cluster computing system for Big Data. It provides high-level APIs in Scala, Java, Python, and R, and an optimized engine that supports general computation graphs for data analysis. It also supports a rich set of higher-level tools including Spark SQL for SQL and DataFrames, MLlib for machine learning, GraphX for graph processing, and Spark Streaming for stream processing.

http://spark.apache.org/

Online Documentation

You can find the latest Spark documentation, including a programming guide, on the project web page. This README file only contains basic setup instructions.

Building Spark

Spark is built using Apache Maven. To build Spark and its example programs, run:

build/mvn -DskipTests clean package

(You do not need to do this if you downloaded a pre-built package.)

You can build Spark using more than one thread by using the -T option with Maven, see "Parallel builds in Maven 3". More detailed documentation is available from the project site, at "Building Spark".

For general development tips, including info on developing Spark using an IDE, see "Useful Developer Tools".

Interactive Scala Shell

The easiest way to start using Spark is through the Scala shell:

./bin/spark-shell

Try the following command, which should return 1000:

scala> sc.parallelize(1 to 1000).count()

Interactive Python Shell

Alternatively, if you prefer Python, you can use the Python shell:

./bin/pyspark

And run the following command, which should also return 1000:

>>> sc.parallelize(range(1000)).count()

Example Programs

Spark also comes with several sample programs in the examples directory. To run one of them, use ./bin/run-example <class> [params]. For example:

./bin/run-example SparkPi

will run the Pi example locally.

You can set the MASTER environment variable when running examples to submit examples to a cluster. This can be a mesos:// or spark:// URL, "yarn" to run on YARN, and "local" to run locally with one thread, or "local[N]" to run locally with N threads. You can also use an abbreviated class name if the class is in the examples package. For instance:

MASTER=spark://host:7077 ./bin/run-example SparkPi

Many of the example programs print usage help if no params are given.

Running Tests

Testing first requires building Spark. Once Spark is built, tests can be run using:

./dev/run-tests

Please see the guidance on how to run tests for a module, or individual tests.

A Note About Hadoop Versions

Spark uses the Hadoop core library to talk to HDFS and other Hadoop-supported storage systems. Because the protocols have changed in different versions of Hadoop, you must build Spark against the same version that your cluster runs.

Please refer to the build documentation at "Specifying the Hadoop Version" for detailed guidance on building for a particular distribution of Hadoop, including building for particular Hive and Hive Thriftserver distributions.

Configuration

Please refer to the Configuration Guide in the online documentation for an overview on how to configure Spark.

Contributing

Please review the Contribution to Spark guide for information on how to get started contributing to the project.