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
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
|
/**
* 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/platform/basic.h"
#include "mongo/db/pipeline/document_source.h"
#include "mongo/db/matcher/expression_algo.h"
#include "mongo/db/pipeline/document_source_match.h"
#include "mongo/db/pipeline/document_source_sequential_document_cache.h"
#include "mongo/db/pipeline/expression_context.h"
#include "mongo/db/pipeline/field_path.h"
#include "mongo/db/pipeline/value.h"
#include "mongo/util/string_map.h"
namespace mongo {
using Parser = DocumentSource::Parser;
using boost::intrusive_ptr;
using std::list;
using std::string;
using std::vector;
DocumentSource::DocumentSource(const intrusive_ptr<ExpressionContext>& pCtx)
: pSource(NULL), pExpCtx(pCtx) {}
namespace {
// Used to keep track of which DocumentSources are registered under which name.
static StringMap<Parser> parserMap;
} // namespace
void DocumentSource::registerParser(string name, Parser parser) {
auto it = parserMap.find(name);
massert(28707,
str::stream() << "Duplicate document source (" << name << ") registered.",
it == parserMap.end());
parserMap[name] = parser;
}
list<intrusive_ptr<DocumentSource>> DocumentSource::parse(
const intrusive_ptr<ExpressionContext>& expCtx, BSONObj stageObj) {
uassert(16435,
"A pipeline stage specification object must contain exactly one field.",
stageObj.nFields() == 1);
BSONElement stageSpec = stageObj.firstElement();
auto stageName = stageSpec.fieldNameStringData();
// Get the registered parser and call that.
auto it = parserMap.find(stageName);
uassert(16436,
str::stream() << "Unrecognized pipeline stage name: '" << stageName << "'",
it != parserMap.end());
return it->second(stageSpec, expCtx);
}
const char* DocumentSource::getSourceName() const {
static const char unknown[] = "[UNKNOWN]";
return unknown;
}
intrusive_ptr<DocumentSource> DocumentSource::optimize() {
return this;
}
namespace {
/**
* Given a set of paths 'dependencies', determines which of those paths will be modified if all
* paths except those in 'preservedPaths' are modified.
*
* For example, extractModifiedDependencies({'a', 'b', 'c.d', 'e'}, {'a', 'b.c', c'}) returns
* {'b', 'e'}, since 'b' and 'e' are not preserved (only 'b.c' is preserved).
*/
std::set<std::string> extractModifiedDependencies(const std::set<std::string>& dependencies,
const std::set<std::string>& preservedPaths) {
std::set<std::string> modifiedDependencies;
// The modified dependencies is *almost* the set difference 'dependencies' - 'preservedPaths',
// except that if p in 'preservedPaths' is a "path prefix" of d in 'dependencies', then 'd'
// should not be included in the modified dependencies.
for (auto&& dependency : dependencies) {
bool preserved = false;
auto firstField = FieldPath::extractFirstFieldFromDottedPath(dependency).toString();
// If even a prefix is preserved, the path is preserved, so search for any prefixes of
// 'dependency' as well. 'preservedPaths' is an *ordered* set, so we only have to search the
// range ['firstField', 'dependency'] to find any prefixes of 'dependency'.
for (auto it = preservedPaths.lower_bound(firstField);
it != preservedPaths.upper_bound(dependency);
++it) {
if (*it == dependency || expression::isPathPrefixOf(*it, dependency)) {
preserved = true;
break;
}
}
if (!preserved) {
modifiedDependencies.insert(dependency);
}
}
return modifiedDependencies;
}
/**
* Returns a pair of pointers to $match stages, either of which can be null. The first entry in the
* pair is a $match stage that can be moved before this stage, the second is a $match stage that
* must remain after this stage.
*/
std::pair<boost::intrusive_ptr<DocumentSourceMatch>, boost::intrusive_ptr<DocumentSourceMatch>>
splitMatchByModifiedFields(const boost::intrusive_ptr<DocumentSourceMatch>& match,
const DocumentSource::GetModPathsReturn& modifiedPathsRet) {
// Attempt to move some or all of this $match before this stage.
std::set<std::string> modifiedPaths;
switch (modifiedPathsRet.type) {
case DocumentSource::GetModPathsReturn::Type::kNotSupported:
// We don't know what paths this stage might modify, so refrain from swapping.
return {nullptr, match};
case DocumentSource::GetModPathsReturn::Type::kAllPaths:
// This stage modifies all paths, so cannot be swapped with a $match at all.
return {nullptr, match};
case DocumentSource::GetModPathsReturn::Type::kFiniteSet:
modifiedPaths = std::move(modifiedPathsRet.paths);
break;
case DocumentSource::GetModPathsReturn::Type::kAllExcept: {
DepsTracker depsTracker;
match->getDependencies(&depsTracker);
auto preservedPaths = modifiedPathsRet.paths;
for (auto&& rename : modifiedPathsRet.renames) {
preservedPaths.insert(rename.first);
}
modifiedPaths = extractModifiedDependencies(depsTracker.fields, preservedPaths);
}
}
return match->splitSourceBy(modifiedPaths, modifiedPathsRet.renames);
}
/**
* If 'pathOfInterest' or some path prefix of 'pathOfInterest' is renamed, returns the new name for
* 'pathOfInterest', otherwise returns boost::none.
* For example, if 'renamedPaths' is {"c.d", "c"}, and 'pathOfInterest' is "c.d.f", returns "c.f".
*/
boost::optional<std::string> findNewName(const StringMap<std::string>& renamedPaths,
std::string pathOfInterest) {
FieldPath fullPathOfInterest(pathOfInterest);
StringBuilder toLookup;
std::size_t pathIndex = 0;
while (pathIndex < fullPathOfInterest.getPathLength()) {
if (pathIndex != 0) {
toLookup << ".";
}
toLookup << fullPathOfInterest.getFieldName(pathIndex++);
auto it = renamedPaths.find(toLookup.stringData());
if (it != renamedPaths.end()) {
const auto& newPathOfPrefix = it->second;
// We found a rename! Note this might be a rename of the prefix of the path, so we have
// to add back on the suffix that was unchanged.
StringBuilder renamedPath;
renamedPath << newPathOfPrefix;
while (pathIndex < fullPathOfInterest.getPathLength()) {
renamedPath << "." << fullPathOfInterest.getFieldName(pathIndex++);
}
return {renamedPath.str()};
}
}
return boost::none;
}
StringMap<std::string> computeNewNamesAssumingAnyPathsNotRenamedAreUnmodified(
const StringMap<std::string>& renamedPaths, const std::set<std::string>& pathsOfInterest) {
StringMap<std::string> renameOut;
for (auto&& ofInterest : pathsOfInterest) {
if (auto newName = findNewName(renamedPaths, ofInterest)) {
renameOut[ofInterest] = *newName;
} else {
// This path was not renamed, assume it was unchanged and map it to itself.
renameOut[ofInterest] = ofInterest;
}
}
return renameOut;
}
} // namespace
Pipeline::SourceContainer::iterator DocumentSource::optimizeAt(
Pipeline::SourceContainer::iterator itr, Pipeline::SourceContainer* container) {
invariant(*itr == this);
// If we are at the end of the pipeline, only optimize in the special case of a cache stage.
if (std::next(itr) == container->end()) {
return dynamic_cast<DocumentSourceSequentialDocumentCache*>(this)
? doOptimizeAt(itr, container)
: container->end();
}
auto nextMatch = dynamic_cast<DocumentSourceMatch*>((*std::next(itr)).get());
if (constraints().canSwapWithMatch && nextMatch && !nextMatch->isTextQuery()) {
// We're allowed to swap with a $match and the stage after us is a $match. Furthermore, the
// $match does not contain a text search predicate, which we do not attempt to optimize
// because such a $match must already be the first stage in the pipeline. We can attempt to
// swap the $match or part of the $match before ourselves.
auto splitMatch = splitMatchByModifiedFields(nextMatch, getModifiedPaths());
invariant(splitMatch.first || splitMatch.second);
if (splitMatch.first) {
// At least part of the $match can be moved before this stage. Erase the original $match
// and put the independent part before this stage. If splitMatch.second is not null,
// then there is a new $match stage to insert after ourselves which is dependent on the
// modified fields.
container->erase(std::next(itr));
container->insert(itr, std::move(splitMatch.first));
if (splitMatch.second) {
container->insert(std::next(itr), std::move(splitMatch.second));
}
// The stage before the new $match may be able to optimize further, if there is such a
// stage.
return std::prev(itr) == container->begin() ? std::prev(itr)
: std::prev(std::prev(itr));
}
}
return doOptimizeAt(itr, container);
}
boost::optional<StringMap<std::string>> DocumentSource::renamedPaths(
const std::set<std::string>& pathsOfInterest) const {
auto modifiedPathsRet = this->getModifiedPaths();
switch (modifiedPathsRet.type) {
case DocumentSource::GetModPathsReturn::Type::kNotSupported:
case DocumentSource::GetModPathsReturn::Type::kAllPaths:
return boost::none;
case DocumentSource::GetModPathsReturn::Type::kFiniteSet: {
for (auto&& modified : modifiedPathsRet.paths) {
for (auto&& ofInterest : pathsOfInterest) {
// Any overlap of the path means the path of interest is not preserved. For
// example, if the path of interest is "a.b", then a modified path of "a",
// "a.b", or "a.b.c" would all signal that "a.b" is not preserved.
if (ofInterest == modified ||
expression::isPathPrefixOf(ofInterest, modified) ||
expression::isPathPrefixOf(modified, ofInterest)) {
// This stage modifies at least one of the fields which the caller is
// interested in, bail out.
return boost::none;
}
}
}
// None of the paths of interest were modified, construct the result map, mapping
// the names after this stage to the names before this stage.
return computeNewNamesAssumingAnyPathsNotRenamedAreUnmodified(modifiedPathsRet.renames,
pathsOfInterest);
}
case DocumentSource::GetModPathsReturn::Type::kAllExcept: {
auto preservedPaths = modifiedPathsRet.paths;
for (auto&& rename : modifiedPathsRet.renames) {
// For the purposes of checking which paths are modified, consider renames to
// preserve the path. We'll circle back later to figure out the new name if
// appropriate.
preservedPaths.insert(rename.first);
}
auto modifiedPaths = extractModifiedDependencies(pathsOfInterest, preservedPaths);
if (modifiedPaths.empty()) {
return computeNewNamesAssumingAnyPathsNotRenamedAreUnmodified(
modifiedPathsRet.renames, pathsOfInterest);
}
return boost::none;
}
}
MONGO_UNREACHABLE;
}
void DocumentSource::serializeToArray(vector<Value>& array,
boost::optional<ExplainOptions::Verbosity> explain) const {
Value entry = serialize(explain);
if (!entry.missing()) {
array.push_back(entry);
}
}
BSONObjSet DocumentSource::allPrefixes(BSONObj obj) {
BSONObjSet out = SimpleBSONObjComparator::kInstance.makeBSONObjSet();
BSONObj last = {};
for (auto&& field : obj) {
BSONObjBuilder builder(last.objsize() + field.size());
builder.appendElements(last);
builder.append(field);
last = builder.obj();
out.insert(last);
}
return out;
}
BSONObjSet DocumentSource::truncateSortSet(const BSONObjSet& sorts,
const std::set<std::string>& fields) {
BSONObjSet out = SimpleBSONObjComparator::kInstance.makeBSONObjSet();
for (auto&& sort : sorts) {
BSONObjBuilder outputSort;
for (auto&& key : sort) {
auto keyName = key.fieldNameStringData();
bool shouldAppend = true;
for (auto&& field : fields) {
if (keyName == field || keyName.startsWith(field + '.')) {
shouldAppend = false;
break;
}
}
if (!shouldAppend) {
break;
}
outputSort.append(key);
}
BSONObj outSortObj = outputSort.obj();
if (!outSortObj.isEmpty()) {
out.insert(outSortObj);
}
}
return out;
}
} // namespace mongo
|