summaryrefslogtreecommitdiff
path: root/src/mongo/db/storage/index_entry_comparison.h
blob: 4fa2cb5d0e857a7a0a3312d905601fee8fa6e308 (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
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
/**
 *    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.
 */

#pragma once

#include <iosfwd>
#include <tuple>
#include <vector>

#include "mongo/bson/simple_bsonobj_comparator.h"
#include "mongo/db/index/index_descriptor.h"
#include "mongo/db/jsobj.h"
#include "mongo/db/namespace_string.h"
#include "mongo/db/record_id.h"
#include "mongo/db/storage/duplicate_key_error_info.h"
#include "mongo/db/storage/key_string.h"

namespace mongo {

/**
 * Represents a single item in an index. An index item simply consists of a key
 * and a disk location.
 */
struct IndexKeyEntry {
    /**
     * Given an index key 'dehyratedKey' with no field names, returns a new BSONObj representing the
     * index key after adding field names according to 'keyPattern'.
     */
    static BSONObj rehydrateKey(const BSONObj& keyPattern, const BSONObj& dehydratedKey) {
        BSONObjBuilder bob;
        BSONObjIterator keyIter(keyPattern);
        BSONObjIterator valueIter(dehydratedKey);

        while (keyIter.more() && valueIter.more()) {
            bob.appendAs(valueIter.next(), keyIter.next().fieldNameStringData());
        }

        invariant(!keyIter.more());
        invariant(!valueIter.more());

        return bob.obj();
    }

    IndexKeyEntry(BSONObj key, RecordId loc) : key(std::move(key)), loc(std::move(loc)) {}

    void serialize(BSONObjBuilder* builder) const {
        builder->append("key"_sd, key);
        loc.serializeToken("RecordId", builder);
    }

    BSONObj key;
    RecordId loc;
};

std::ostream& operator<<(std::ostream& stream, const IndexKeyEntry& entry);

inline bool operator==(const IndexKeyEntry& lhs, const IndexKeyEntry& rhs) {
    return SimpleBSONObjComparator::kInstance.evaluate(lhs.key == rhs.key) && (lhs.loc == rhs.loc);
}

inline bool operator!=(const IndexKeyEntry& lhs, const IndexKeyEntry& rhs) {
    return !(lhs == rhs);
}

/**
 * Represents KeyString struct containing a KeyString::Value and its RecordId
 */
struct KeyStringEntry {
    KeyStringEntry(KeyString::Value ks, RecordId id) : keyString(ks), loc(std::move(id)) {
        if (!kDebugBuild) {
            return;
        }
        loc.withFormat(
            [](RecordId::Null n) { invariant(false); },
            [&](int64_t rid) {
                invariant(loc == KeyString::decodeRecordIdLongAtEnd(ks.getBuffer(), ks.getSize()));
            },
            [&](const char* str, int size) {
                invariant(loc == KeyString::decodeRecordIdStrAtEnd(ks.getBuffer(), ks.getSize()));
            });
    }

    KeyString::Value keyString;
    RecordId loc;
};

/**
 * Describes a query that can be compared against an IndexKeyEntry in a way that allows
 * expressing exclusiveness on a prefix of the key. This is mostly used to express a location to
 * seek to in an index that may not be representable as a valid key.
 *
 * If 'firstExclusive' is negative, the "key" used for comparison is the concatenation of the first
 * 'prefixLen' elements of 'keyPrefix' followed by the last 'keySuffix.size() - prefixLen' elements
 * of 'keySuffix'.
 *
 * The comparison is exclusive if 'firstExclusive' is non-negative, and any portion of the key after
 * that index will be omitted. The value of 'firstExclusive' must be at least 'prefixLen' - 1.
 *
 * e.g.
 *
 *  Suppose that
 *
 *      keyPrefix = { "" : 1, "" : 2 }
 *      prefixLen = 1
 *      keySuffix = [ IGNORED, { "" : 5 }, { "" : 9 } ]
 *      firstExclusive = 1
 *
 *      ==> key is { "" : 1, "" : 5 }
 *          with the comparison being done exclusively
 *
 *  Suppose that
 *
 *      keyPrefix = { "" : 1, "" : 2 }
 *      prefixLen = 1
 *      keySuffix = IGNORED
 *      firstExclusive = 0
 *
 *      ==> represented key is { "" : 1 }
 *          with the comparison being done exclusively
 *
 *  Suppose that
 *
 *      keyPrefix = { "" : 1, "" : 2 }
 *      prefixLen = 1
 *      keySuffix = [ IGNORED, { "" : 5 }, { "" : 9 } ]
 *      firstExclusive = -1
 *
 *      ==> key is { "" : 1, "" : 5, "" : 9 }
 *          with the comparison being done inclusively
 */
struct IndexSeekPoint {
    BSONObj keyPrefix;

    /**
     * Use this many fields in 'keyPrefix'.
     */
    int prefixLen = 0;

    /**
     * Elements starting at index 'prefixLen' are logically appended to the prefix.
     * The elements before index 'prefixLen' should be ignored.
     */
    std::vector<const BSONElement*> keySuffix;

    /**
     * If non-negative, then the comparison will be exclusive and any elements after index
     * 'firstExclusive' are ignored. Otherwise all elements are considered and the comparison will
     * be inclusive.
     */
    int firstExclusive = -1;
};

/**
 * Compares two different IndexKeyEntry instances. The existence of compound indexes necessitates
 * some complicated logic. This is meant to support the comparisons of IndexKeyEntries (that are
 * stored in an index) with IndexSeekPoints to support fine-grained control over whether the ranges
 * of various keys comprising a compound index are inclusive or exclusive.
 */
class IndexEntryComparison {
public:
    IndexEntryComparison(Ordering order) : _order(order) {}

    bool operator()(const IndexKeyEntry& lhs, const IndexKeyEntry& rhs) const;

    /**
     * Compares two IndexKeyEntries and returns -1 if lhs < rhs, 1 if lhs > rhs, and 0
     * otherwise.
     *
     * IndexKeyEntries are compared lexicographically field by field in the BSONObj, followed by
     * the RecordId.
     */
    int compare(const IndexKeyEntry& lhs, const IndexKeyEntry& rhs) const;

    /**
     * Encodes the SeekPoint into a KeyString object suitable to pass in to seek().
     *
     * A KeyString is used for seeking an iterator to a position in a sorted index. The difference
     * between a query KeyString and the KeyStrings inserted into indexes is that a query KeyString
     * can have an exclusive discriminator, which forces the key to compare less than or greater to
     * a matching key in the index. This means that the first matching key is not returned, but
     * rather the one immediately after. The meaning of "after" depends on the cursor directory,
     * isForward.
     *
     * If a field is marked as exclusive, then comparisons stop after that field and return
     * either higher or lower, even if that field compares equal. If firstExclusive is non-negative,
     * then the field at the corresponding index is marked as exclusive and any subsequent fields
     * are ignored.
     *
     * Returned objects are for use in lookups only and should never be inserted into the
     * database, as their format may change. The only reason this is the same type as the
     * entries in an index is to support storage engines that require comparators that take
     * arguments of the same type.
     */
    static KeyString::Value makeKeyStringFromSeekPointForSeek(const IndexSeekPoint& seekPoint,
                                                              KeyString::Version version,
                                                              Ordering ord,
                                                              bool isForward);

    /**
     * Encodes the BSON Key into a KeyString object to pass in to SortedDataInterface::seek().
     *
     * `isForward` and `inclusive` together decide which discriminator we will put into the
     * KeyString. This logic is closely related to how WiredTiger uses its API
     * (search_near/prev/next) to do the seek. Other storage engines' SortedDataInterface should use
     * the discriminator to deduce the `inclusive` and the use their own ways to seek to the right
     * position.
     *
     * 1. When isForward == true, inclusive == true, bsonKey will be encoded with kExclusiveBefore
     * (which is less than bsonKey). WT's search_near() could land either on the previous key or
     * bsonKey. WT will selectively call next() if it's on the previous key.
     *
     * 2. When isForward == true, inclusive == false, bsonKey will be encoded with kExclusiveAfter
     * (which is greater than bsonKey). WT's search_near() could land either on bsonKey or the next
     * key. WT will selectively call next() if it's on bsonKey.
     *
     * 3. When isForward == false, inclusive == true, bsonKey will be encoded with kExclusiveAfter
     * (which is greater than bsonKey). WT's search_near() could land either on bsonKey or the next
     * key. WT will selectively call prev() if it's on the next key.
     *
     * 4. When isForward == false, inclusive == false, bsonKey will be encoded with kExclusiveBefore
     * (which is less than bsonKey). WT's search_near() could land either on the previous key or the
     * bsonKey. WT will selectively call prev() if it's on bsonKey.
     */
    static KeyString::Value makeKeyStringFromBSONKeyForSeek(const BSONObj& bsonKey,
                                                            KeyString::Version version,
                                                            Ordering ord,
                                                            bool isForward,
                                                            bool inclusive);

    /**
     * Encodes the BSON Key into a KeyString object to pass in to SortedDataInterface::seek()
     * or SortedDataInterface::setEndPosition().
     *
     * This funcition is similar to IndexEntryComparison::makeKeyStringFromBSONKeyForSeek()
     * but allows you to pick your own KeyString::Discriminator based on wether or not the
     * resulting KeyString is for the start key or end key of a seek.
     */
    static KeyString::Value makeKeyStringFromBSONKey(const BSONObj& bsonKey,
                                                     KeyString::Version version,
                                                     Ordering ord,
                                                     KeyString::Discriminator discrim);

private:
    // Ordering is used in comparison() to compare BSONElements
    const Ordering _order;

};  // struct IndexEntryComparison

/**
 * Returns the formatted error status about the duplicate key.
 */
Status buildDupKeyErrorStatus(const BSONObj& key,
                              const NamespaceString& collectionNamespace,
                              const std::string& indexName,
                              const BSONObj& keyPattern,
                              const BSONObj& indexCollation,
                              DuplicateKeyErrorInfo::FoundValue&& foundValue = stdx::monostate{},
                              boost::optional<RecordId> duplicateRid = boost::none);

Status buildDupKeyErrorStatus(const KeyString::Value& keyString,
                              const NamespaceString& collectionNamespace,
                              const std::string& indexName,
                              const BSONObj& keyPattern,
                              const BSONObj& indexCollation,
                              const Ordering& ordering);

Status buildDupKeyErrorStatus(OperationContext* opCtx,
                              const BSONObj& key,
                              const IndexDescriptor* desc);

Status buildDupKeyErrorStatus(OperationContext* opCtx,
                              const KeyString::Value& keyString,
                              const Ordering& ordering,
                              const IndexDescriptor* desc);

}  // namespace mongo