summaryrefslogtreecommitdiff
path: root/src/mongo/db/exec/and_sorted.cpp
blob: 5d35a7a1bd2456f60d1fcfd987b92c18670c5bb6 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
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
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
/**
 *    Copyright (C) 2018-present MongoDB, Inc.
 *
 *    This program is free software: you can redistribute it and/or modify
 *    it under the terms of the Server Side Public License, version 1,
 *    as published by MongoDB, Inc.
 *
 *    This program is distributed in the hope that it will be useful,
 *    but WITHOUT ANY WARRANTY; without even the implied warranty of
 *    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *    Server Side Public License for more details.
 *
 *    You should have received a copy of the Server Side Public License
 *    along with this program. If not, see
 *    <http://www.mongodb.com/licensing/server-side-public-license>.
 *
 *    As a special exception, the copyright holders give permission to link the
 *    code of portions of this program with the OpenSSL library under certain
 *    conditions as described in each individual source file and distribute
 *    linked combinations including the program with the OpenSSL library. You
 *    must comply with the Server Side Public License in all respects for
 *    all of the code used other than as permitted herein. If you modify file(s)
 *    with this exception, you may extend this exception to your version of the
 *    file(s), but you are not obligated to do so. If you do not wish to do so,
 *    delete this exception statement from your version. If you delete this
 *    exception statement from all source files in the program, then also delete
 *    it in the license file.
 */

#include "mongo/db/exec/and_sorted.h"

#include <memory>

#include "mongo/db/exec/and_common.h"
#include "mongo/db/exec/scoped_timer.h"
#include "mongo/db/exec/working_set_common.h"
#include "mongo/util/str.h"

namespace mongo {

using std::numeric_limits;
using std::unique_ptr;
using std::vector;

// static
const char* AndSortedStage::kStageType = "AND_SORTED";

AndSortedStage::AndSortedStage(ExpressionContext* expCtx, WorkingSet* ws)
    : PlanStage(kStageType, expCtx),
      _ws(ws),
      _targetNode(numeric_limits<size_t>::max()),
      _targetId(WorkingSet::INVALID_ID),
      _isEOF(false) {}


void AndSortedStage::addChild(std::unique_ptr<PlanStage> child) {
    _children.emplace_back(std::move(child));
}

bool AndSortedStage::isEOF() {
    return _isEOF;
}

PlanStage::StageState AndSortedStage::doWork(WorkingSetID* out) {
    if (isEOF()) {
        return PlanStage::IS_EOF;
    }

    if (0 == _specificStats.failedAnd.size()) {
        _specificStats.failedAnd.resize(_children.size());
    }

    // If we don't have any nodes that we're work()-ing until they hit a certain RecordId...
    if (0 == _workingTowardRep.size()) {
        // Get a target RecordId.
        return getTargetRecordId(out);
    }

    // Move nodes toward the target RecordId.
    // If all nodes reach the target RecordId, return it.  The next call to work() will set a new
    // target.
    return moveTowardTargetRecordId(out);
}

PlanStage::StageState AndSortedStage::getTargetRecordId(WorkingSetID* out) {
    verify(numeric_limits<size_t>::max() == _targetNode);
    verify(WorkingSet::INVALID_ID == _targetId);
    verify(RecordId() == _targetRecordId);

    // Pick one, and get a RecordId to work toward.
    WorkingSetID id = WorkingSet::INVALID_ID;
    StageState state = _children[0]->work(&id);

    if (PlanStage::ADVANCED == state) {
        WorkingSetMember* member = _ws->get(id);

        // The child must give us a WorkingSetMember with a record id, since we intersect index keys
        // based on the record id. The planner ensures that the child stage can never produce an WSM
        // with no record id.
        invariant(member->hasRecordId());

        // We have a value from one child to AND with.
        _targetNode = 0;
        _targetId = id;
        _targetRecordId = member->recordId;

        // Ensure that the BSONObj underlying the WorkingSetMember is owned in case we yield.
        member->makeObjOwnedIfNeeded();

        // We have to AND with all other children.
        for (size_t i = 1; i < _children.size(); ++i) {
            _workingTowardRep.push(i);
        }

        return PlanStage::NEED_TIME;
    } else if (PlanStage::IS_EOF == state) {
        _isEOF = true;
        return state;
    } else {
        if (PlanStage::NEED_YIELD == state) {
            *out = id;
        }

        // NEED_TIME, NEED_YIELD.
        return state;
    }
}

PlanStage::StageState AndSortedStage::moveTowardTargetRecordId(WorkingSetID* out) {
    verify(numeric_limits<size_t>::max() != _targetNode);
    verify(WorkingSet::INVALID_ID != _targetId);

    // We have nodes that haven't hit _targetRecordId yet.
    size_t workingChildNumber = _workingTowardRep.front();
    auto& next = _children[workingChildNumber];
    WorkingSetID id = WorkingSet::INVALID_ID;
    StageState state = next->work(&id);

    if (PlanStage::ADVANCED == state) {
        WorkingSetMember* member = _ws->get(id);

        // The child must give us a WorkingSetMember with a record id, since we intersect index keys
        // based on the record id. The planner ensures that the child stage can never produce an WSM
        // with no record id.
        invariant(member->hasRecordId());

        if (member->recordId == _targetRecordId) {
            // The front element has hit _targetRecordId.  Don't move it forward anymore/work on
            // another element.
            _workingTowardRep.pop();
            AndCommon::mergeFrom(_ws, _targetId, *member);
            _ws->free(id);

            if (0 == _workingTowardRep.size()) {
                WorkingSetID toReturn = _targetId;

                _targetNode = numeric_limits<size_t>::max();
                _targetId = WorkingSet::INVALID_ID;
                _targetRecordId = RecordId();

                *out = toReturn;
                return PlanStage::ADVANCED;
            }
            // More children need to be advanced to _targetRecordId.
            return PlanStage::NEED_TIME;
        } else if (member->recordId < _targetRecordId) {
            // The front element of _workingTowardRep hasn't hit the thing we're AND-ing with
            // yet.  Try again later.
            _ws->free(id);
            return PlanStage::NEED_TIME;
        } else {
            // member->recordId > _targetRecordId.
            // _targetRecordId wasn't successfully AND-ed with the other sub-plans.  We toss it and
            // try AND-ing with the next value.
            _specificStats.failedAnd[_targetNode]++;

            _ws->free(_targetId);
            _targetNode = workingChildNumber;
            _targetRecordId = member->recordId;
            _targetId = id;

            // Ensure that the BSONObj underlying the WorkingSetMember is owned in case we yield.
            member->makeObjOwnedIfNeeded();

            _workingTowardRep = std::queue<size_t>();
            for (size_t i = 0; i < _children.size(); ++i) {
                if (workingChildNumber != i) {
                    _workingTowardRep.push(i);
                }
            }
            // Need time to chase after the new _targetRecordId.
            return PlanStage::NEED_TIME;
        }
    } else if (PlanStage::IS_EOF == state) {
        _isEOF = true;
        _ws->free(_targetId);
        return state;
    } else {
        if (PlanStage::NEED_YIELD == state) {
            *out = id;
        }

        return state;
    }
}

unique_ptr<PlanStageStats> AndSortedStage::getStats() {
    _commonStats.isEOF = isEOF();

    unique_ptr<PlanStageStats> ret =
        std::make_unique<PlanStageStats>(_commonStats, STAGE_AND_SORTED);
    ret->specific = std::make_unique<AndSortedStats>(_specificStats);
    for (size_t i = 0; i < _children.size(); ++i) {
        ret->children.emplace_back(_children[i]->getStats());
    }

    return ret;
}

const SpecificStats* AndSortedStage::getSpecificStats() const {
    return &_specificStats;
}

}  // namespace mongo