Package org.apache.pig.newplan.logical.relational

Examples of org.apache.pig.newplan.logical.relational.LogicalSchema

532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
                    if (!((ProjectExpression)op).isProjectStar()) {
                        long uid = ((ProjectExpression)op).getFieldSchema().uid;
                        uids.add(uid);
                    } else {
                        LogicalRelationalOperator ref = ((ProjectExpression)op).findReferent();
                        LogicalSchema s = ref.getSchema();
                        if (s == null) {
                            throw new SchemaNotDefinedException("Schema not defined for " + ref.getAlias());
                        }
                        for(LogicalFieldSchema f: s.getFields()) {
                            uids.add(f.uid);
                        }
                    }
                }
            }
View Full Code Here
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
        private Set<Long> setOutputUids(LogicalRelationalOperator op) throws FrontendException {
           
            List<Operator> ll = plan.getSuccessors(op);
            Set<Long> uids = new HashSet<Long>();
           
            LogicalSchema s = op.getSchema();
            if (s == null) {
                throw new SchemaNotDefinedException("Schema for " + op.getName() + " is not defined.");
            }
                           
            if (ll != null) {
                // if this is not sink, the output uids are union of input uids of its successors
                for(Operator succ: ll) {
                    Set<Long> inputUids = (Set<Long>)succ.getAnnotation(INPUTUIDS);
                    if (inputUids != null) {
                        Iterator<Long> iter = inputUids.iterator();
                        while(iter.hasNext()) {
                            long uid = iter.next();
                           
                            if (s.findField(uid) != -1) {
                                uids.add(uid);
                            }
                        }
                    }
                }
            } else {
                // if  it's leaf, set to its schema               
                for(int i=0; i<s.size(); i++) {
                    uids.add(s.getField(i).uid);
                }                               
            }
           
            op.annotate(OUTPUTUIDS, uids);
            return uids;
View Full Code Here
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
                //predecessors haven't been setup, return null
                return null;
            }
        }

        LogicalSchema inputSchema = new LogicalSchema();
        List<Operator> succs = plan.getSuccessors(this);

        if (succs!=null) {
            for(Operator lo : succs){
                if (((LogicalExpression)lo).getFieldSchema()==null) {
                    inputSchema = null;
                    break;
                }
                inputSchema.addField(((LogicalExpression)lo).getFieldSchema());
            }
        }

        // Since ef only set one time, we never change its value, so we can optimize it by instantiate only once.
        // This significantly optimize the performance of frontend (PIG-1738)
View Full Code Here
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
        // store D into 'whatever';
       
        // A = load
        LogicalPlan lp = new LogicalPlan();
        {
          LogicalSchema aschema = new LogicalSchema();
          aschema.addField(new LogicalSchema.LogicalFieldSchema(
              "x", null, DataType.BYTEARRAY));
          aschema.addField(new LogicalSchema.LogicalFieldSchema(
              "y", null, DataType.BYTEARRAY));
          LOLoad A = new LOLoad(new FileSpec("bla", new FuncSpec("PigStorage", "\t")), aschema, lp, null);
          A.setAlias("A");
          lp.add(A);
         
          // B = load
          LogicalSchema bschema = new LogicalSchema();
          bschema.addField(new LogicalSchema.LogicalFieldSchema(
              "a", null, DataType.BYTEARRAY));
          bschema.addField(new LogicalSchema.LogicalFieldSchema(
              "b", null, DataType.BYTEARRAY));
          LOLoad B = new LOLoad(new FileSpec("morebla", new FuncSpec("PigStorage", "\t")), bschema, lp, null);
          B.setAlias("B");
          lp.add(B);
         
          // C = join
          LogicalSchema cschema = new LogicalSchema();
          cschema.addField(new LogicalSchema.LogicalFieldSchema(
              "x", null, DataType.BYTEARRAY));
          cschema.addField(new LogicalSchema.LogicalFieldSchema(
              "y", null, DataType.BYTEARRAY));
          cschema.addField(new LogicalSchema.LogicalFieldSchema(
              "a", null, DataType.BYTEARRAY));
          cschema.addField(new LogicalSchema.LogicalFieldSchema(
              "b", null, DataType.BYTEARRAY));
          MultiMap<Integer, LogicalExpressionPlan> mm =
                new MultiMap<Integer, LogicalExpressionPlan>();
          LogicalExpressionPlan aprojplan = new LogicalExpressionPlan();
          LOJoin C = new LOJoin(lp, mm, JOINTYPE.HASH, new boolean[] {true, true});
          new ProjectExpression(aprojplan, 0, 0, C);
          LogicalExpressionPlan bprojplan = new LogicalExpressionPlan();
          new ProjectExpression(bprojplan, 1, 0, C);
          mm.put(0, aprojplan);
          mm.put(1, bprojplan);
         
          C.setAlias("C");
          lp.add(C);
          lp.connect(A, C);
          lp.connect(B, C);
       
          // D = filter
          LogicalExpressionPlan filterPlan = new LogicalExpressionPlan();
          LOFilter D = new LOFilter(lp, filterPlan);
          ProjectExpression fx = new ProjectExpression(filterPlan, 0, 0, D);
          ConstantExpression fc0 = new ConstantExpression(filterPlan, new Integer(0), new LogicalFieldSchema(null, null, DataType.BYTEARRAY));
          EqualExpression eq1 = new EqualExpression(filterPlan, fx, fc0);
          ProjectExpression fanotherx = new ProjectExpression(filterPlan, 0, 0, D);
          ProjectExpression fa = new ProjectExpression(filterPlan, 0, 2, D);
          EqualExpression eq2 = new EqualExpression(filterPlan, fanotherx, fa);
          AndExpression and1 = new AndExpression(filterPlan, eq1, eq2);
          ProjectExpression fb = new ProjectExpression(filterPlan, 0, 3, D);
          ConstantExpression fc1 = new ConstantExpression(filterPlan, new Integer(1),new LogicalFieldSchema(null, null, DataType.BYTEARRAY));
          EqualExpression eq3 = new EqualExpression(filterPlan, fb, fc1);
          AndExpression and2 = new AndExpression(filterPlan, and1, eq3);
          ProjectExpression fanotherb = new ProjectExpression(filterPlan, 0, 3, D);
          ProjectExpression fy = new ProjectExpression(filterPlan, 0, 1, D);
          EqualExpression eq4 = new EqualExpression(filterPlan, fy, fanotherb);
          new AndExpression(filterPlan, and2, eq4);
         
          D.setAlias("D");
          // Connect D to B, since the transform has happened.
          lp.add(D);
          lp.connect(C, D);
        }
       
        System.out.println(lp);
        LogicalPlanOptimizer optimizer = new LogicalPlanOptimizer(lp, 500, null);
        optimizer.optimize();
       
        LogicalPlan expected = new LogicalPlan();
        {
            // A = load
          LogicalSchema aschema = new LogicalSchema();
          aschema.addField(new LogicalSchema.LogicalFieldSchema(
              "x", null, DataType.BYTEARRAY));
          aschema.addField(new LogicalSchema.LogicalFieldSchema(
              "y", null, DataType.BYTEARRAY));
          LOLoad A = new LOLoad(new FileSpec("bla", new FuncSpec("PigStorage", "\t")), aschema, expected, null);
          expected.add(A);
         
          // DA = filter
          LogicalExpressionPlan DAfilterPlan = new LogicalExpressionPlan();
          LOFilter DA = new LOFilter(expected, DAfilterPlan);
          ProjectExpression fx = new ProjectExpression(DAfilterPlan, 0, 0, DA);
          fx.neverUseForRealSetFieldSchema(new LogicalFieldSchema(null, null, DataType.BYTEARRAY));
          ConstantExpression fc0 = new ConstantExpression(DAfilterPlan, new Integer(0), new LogicalFieldSchema(null, null, DataType.BYTEARRAY));
          new EqualExpression(DAfilterPlan, fx, fc0);
         
          DA.neverUseForRealSetSchema(aschema);
          expected.add(DA);
          expected.connect(A, DA);
         
          // B = load
          LogicalSchema bschema = new LogicalSchema();
          bschema.addField(new LogicalSchema.LogicalFieldSchema(
              "a", null, DataType.BYTEARRAY));
          bschema.addField(new LogicalSchema.LogicalFieldSchema(
              "b", null, DataType.BYTEARRAY));
          LOLoad B = new LOLoad(new FileSpec("morebla", new FuncSpec("PigStorage", "\t")), bschema, expected, null);
          expected.add(B);
         
          // DB = filter
          LogicalExpressionPlan DBfilterPlan = new LogicalExpressionPlan();
            LOFilter DB = new LOFilter(expected, DBfilterPlan);
          ProjectExpression fb = new ProjectExpression(DBfilterPlan, 0, 1, DB);
          fb.neverUseForRealSetFieldSchema(new LogicalFieldSchema(null, null, DataType.BYTEARRAY));
          ConstantExpression fc1 = new ConstantExpression(DBfilterPlan, new Integer(1), new LogicalFieldSchema(null, null, DataType.BYTEARRAY));
          new EqualExpression(DBfilterPlan, fb, fc1);

          DB.neverUseForRealSetSchema(bschema);
          expected.add(DB);
          expected.connect(B, DB);
         
          // C = join
          LogicalSchema cschema = new LogicalSchema();
          cschema.addField(new LogicalSchema.LogicalFieldSchema(
              "D::x", null, DataType.BYTEARRAY));
          cschema.addField(new LogicalSchema.LogicalFieldSchema(
              "D::y", null, DataType.BYTEARRAY));
          cschema.addField(new LogicalSchema.LogicalFieldSchema(
              "null::a", null, DataType.BYTEARRAY));
          cschema.addField(new LogicalSchema.LogicalFieldSchema(
              "null::b", null, DataType.BYTEARRAY));
          cschema.getField(0).uid = 1;
          cschema.getField(1).uid = 2;
          cschema.getField(2).uid = 3;
          cschema.getField(3).uid = 4;
          LogicalExpressionPlan aprojplan = new LogicalExpressionPlan();
            MultiMap<Integer, LogicalExpressionPlan> mm =
                new MultiMap<Integer, LogicalExpressionPlan>();
            LOJoin C = new LOJoin(expected, mm, JOINTYPE.HASH, new boolean[] {true, true});
View Full Code Here
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
                    op = iter.next();
                    if( op instanceof ProjectExpression ) {
                        ProjectExpression proj = (ProjectExpression)op;
                        if( proj.isProjectStar() ) {
                            LogicalRelationalOperator pred = (LogicalRelationalOperator)filter.getPlan().getPredecessors(filter).get(0);
                            LogicalSchema predSchema = pred.getSchema();
                            if (predSchema!=null) {
                                for (int i=0;i<predSchema.size();i++) {
                                    uids.add(predSchema.getField(i).uid);
                                    types.add(predSchema.getField(i).type);
                                }
                            }
                        } else {
                            uids.add(proj.getFieldSchema().uid);
                            types.add(proj.getFieldSchema().type);
View Full Code Here
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
         * @param uids Uids to check for
         * @return true if given LogicalRelationalOperator has all the given uids
         */
        private boolean hasAll(LogicalRelationalOperator op, Pair<List<Long>,
                List<Byte>> uidWithTypes) throws FrontendException {
            LogicalSchema schema = op.getSchema();
           
            if (schema==null)
                return false;
           
            List<Long> uids = uidWithTypes.first;
            List<Byte> types = uidWithTypes.second;
           
            for (int i=0;i<uids.size();i++) {
                boolean found = false;
                for (LogicalSchema.LogicalFieldSchema fs : schema.getFields()) {
                    if (fs.uid==uids.get(i) && fs.type==types.get(i))
                        found = true;
                }
                if (!found)
                    return false;
View Full Code Here
301
302
303
304
305
306
307
308
309
310
311
312
313
            }
        }
       
        // check if a relational operator contains all of the specified uids
        private boolean hasAll(LogicalRelationalOperator op, Set<Long> uids) throws FrontendException {
            LogicalSchema schema = op.getSchema();
            for(long uid: uids) {
                if (schema.findField(uid) == -1) {
                    return false;
                }
            }
           
            return true;
View Full Code Here
73
74
75
76
77
78
79
80
81
82
83
84
85
86
                return;
            }
            else {
                // Get the schema for this operator and search it for the matching uid
                int match = -1;
                LogicalSchema schema = pred.getSchema();
                if (schema==null)
                    return;
                List<LogicalSchema.LogicalFieldSchema> fields = schema.getFields();
                for (int i = 0; i < fields.size(); i++) {
                    if (fields.get(i).uid == myUid) {
                        match = i;
                        break;
                    }
View Full Code Here
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
       
        // First check if we have a load with a map in it or not
        List<Operator> sources = currentPlan.getSources();
       
        for( Operator source : sources ) {
            LogicalSchema schema = ((LogicalRelationalOperator)source).getSchema();
            // If any of the loads has a null schema we dont know the ramifications here
            // so we skip this optimization
            if( schema == null ) {
                return false;
            }
        }
                   
        // Now we check what keys are needed
        MapMarker marker = new MapMarker(currentPlan);
        marker.visit();
       
        // If the uid is the input uid of LOStore, LOCogroup, LOUnion, UserFunc, that means
        // the entire map may be used. For simplicity, we do not prune any map key in this case
        Set<Long> fullMapUids = new HashSet<Long>();
        FullMapCollector collector = new FullMapCollector(currentPlan, fullMapUids);
        collector.visit();
       
        // If we have found specific keys which are needed then we return true;
        // Else if we dont have any specific keys we return false
        boolean hasAnnotation = false;
        for( Operator source : sources ) {
            Map<Integer,Set<String>> annotationValue =
                (Map<Integer, Set<String>>) ((LogicalRelationalOperator)source).getAnnotation(REQUIRED_MAPKEYS);
           
            // Now for all full maps found in sinks we cannot prune them at source
            if( ! fullMapUids.isEmpty() && annotationValue != null &&
                    !annotationValue.isEmpty() ) {
                Integer[] annotationKeyArray = annotationValue.keySet().toArray( new Integer[0] );
                LogicalSchema sourceSchema = ((LogicalRelationalOperator)source).getSchema();
                for( Integer col : annotationKeyArray ) {                 
                    if( fullMapUids.contains(sourceSchema.getField(col).uid)) {
                        annotationValue.remove( col );
                    }
                }
            }
           
View Full Code Here
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
    public LogicalSchema.LogicalFieldSchema getFieldSchema() throws FrontendException {
        if (fieldSchema!=null)
            return fieldSchema;
        LogicalRelationalOperator referent = findReferent();
       
        LogicalSchema schema = referent.getSchema();
       
        if (attachedRelationalOp instanceof LOGenerate && plan.getSuccessors(this)==null) {
            if (!(findReferent() instanceof LOInnerLoad)||
                    ((LOInnerLoad)findReferent()).sourceIsBag()) {
                String alias = findReferent().getAlias();
                List<LOInnerLoad> innerLoads = LOForEach.findReacheableInnerLoadFromBoundaryProject(this);
               
                // pull tuple information from innerload
                if (innerLoads.get(0).getProjection().getFieldSchema().schema!=null &&
                        innerLoads.get(0).getProjection().getFieldSchema().schema.isTwoLevelAccessRequired()) {
                    LogicalFieldSchema originalTupleFieldSchema = innerLoads.get(0).getProjection().getFieldSchema().schema.getField(0);
                    LogicalFieldSchema newTupleFieldSchema = new LogicalFieldSchema(originalTupleFieldSchema.alias,
                            schema, DataType.TUPLE);
                    newTupleFieldSchema.uid = originalTupleFieldSchema.uid;
                    LogicalSchema newTupleSchema = new LogicalSchema();
                    newTupleSchema.setTwoLevelAccessRequired(true);
                    newTupleSchema.addField(newTupleFieldSchema);
                    fieldSchema = new LogicalSchema.LogicalFieldSchema(alias, newTupleSchema, DataType.BAG);
                }
                else {
                    fieldSchema = new LogicalSchema.LogicalFieldSchema(alias, schema, DataType.BAG);
                }
                fieldSchema.uid = innerLoads.get(0).getProjection().getFieldSchema().uid;
            }
            else {
                if (findReferent().getSchema()!=null)
                    fieldSchema = findReferent().getSchema().getField(0);
            }
            if (fieldSchema!=null)
                uidOnlyFieldSchema = fieldSchema.mergeUid(uidOnlyFieldSchema);
        }
        else {
            if (schema == null) {
                fieldSchema = new LogicalSchema.LogicalFieldSchema(null, null, DataType.BYTEARRAY);
                uidOnlyFieldSchema = fieldSchema.mergeUid(uidOnlyFieldSchema);
            }
            else {
                int index = -1;
                if (!isProjectStar() && uidOnlyFieldSchema!=null) {
                    long uid = uidOnlyFieldSchema.uid;
                    for (int i=0;i<schema.size();i++) {
                        LogicalFieldSchema fs = schema.getField(i);
                        if (fs.uid==uid) {
                            index = i;
                        }
                    }
                }
                if (index==-1)
                    index = col;
               
                if (!isProjectStar()) {
                    if (schema!=null && schema.size()>index)
                        fieldSchema = schema.getField(index);
                    else
                        fieldSchema = new LogicalSchema.LogicalFieldSchema(null, null, DataType.BYTEARRAY);
                    uidOnlyFieldSchema = fieldSchema.cloneUid();
                }
                else {
                    LogicalSchema newTupleSchema = null;
                    if (schema!=null)
                        newTupleSchema = schema.deepCopy();
                    fieldSchema = new LogicalSchema.LogicalFieldSchema(null, newTupleSchema, DataType.TUPLE);
                    uidOnlyFieldSchema = fieldSchema.mergeUid(uidOnlyFieldSchema);
                }
View Full Code Here
TOP

Related Classes of org.apache.pig.newplan.logical.relational.LogicalSchema

  • org.apache.pig.impl.util.Utils
  • org.apache.pig.newplan.logical.expression.DereferenceExpression
  • org.apache.pig.newplan.logical.expression.ProjectExpression
  • org.apache.pig.newplan.logical.expression.UserFuncExpression
  • org.apache.pig.newplan.logical.LogicalPlanMigrationVistor
  • org.apache.pig.newplan.logical.optimizer.ProjectionPatcher$ProjectionRewriter
  • org.apache.pig.newplan.logical.rules.AddForEach$AddForEachTransformer
  • org.apache.pig.newplan.logical.rules.ColumnPruneHelper
  • org.apache.pig.newplan.logical.rules.ColumnPruneHelper$ColumnDependencyVisitor
  • org.apache.pig.newplan.logical.rules.ColumnPruneVisitor
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.