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
|
/**
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "SchemaEvolution.hh"
#include "orc/Exceptions.hh"
namespace orc {
SchemaEvolution::SchemaEvolution(const std::shared_ptr<Type>& _readType, const Type* fileType)
: readType(_readType) {
if (readType) {
buildConversion(readType.get(), fileType);
} else {
for (uint64_t i = 0; i <= fileType->getMaximumColumnId(); ++i) {
safePPDConversionMap.insert(i);
}
}
}
const Type* SchemaEvolution::getReadType(const Type& fileType) const {
auto ret = readTypeMap.find(fileType.getColumnId());
return ret == readTypeMap.cend() ? &fileType : ret->second;
}
inline void invalidConversion(const Type* readType, const Type* fileType) {
throw SchemaEvolutionError("Cannot convert from " + fileType->toString() + " to " +
readType->toString());
}
struct EnumClassHash {
template <typename T>
std::size_t operator()(T t) const {
return static_cast<std::size_t>(t);
}
};
bool isNumeric(const Type& type) {
auto kind = type.getKind();
return kind == BOOLEAN || kind == BYTE || kind == SHORT || kind == INT || kind == LONG ||
kind == FLOAT || kind == DOUBLE;
}
bool isStringVariant(const Type& type) {
auto kind = type.getKind();
return kind == STRING || kind == CHAR || kind == VARCHAR;
}
bool isDecimal(const Type& type) {
auto kind = type.getKind();
return kind == DECIMAL;
}
bool isTimestamp(const Type& type) {
auto kind = type.getKind();
return kind == TIMESTAMP || kind == TIMESTAMP_INSTANT;
}
struct ConversionCheckResult {
bool isValid;
bool needConvert;
};
ConversionCheckResult checkConversion(const Type& readType, const Type& fileType) {
ConversionCheckResult ret = {false, false};
if (readType.getKind() == fileType.getKind()) {
ret.isValid = true;
if (fileType.getKind() == CHAR || fileType.getKind() == VARCHAR) {
ret.isValid = readType.getMaximumLength() == fileType.getMaximumLength();
} else if (fileType.getKind() == DECIMAL) {
ret.needConvert = readType.getPrecision() != fileType.getPrecision() ||
readType.getScale() != fileType.getScale();
}
} else {
switch (fileType.getKind()) {
case BOOLEAN:
case BYTE:
case SHORT:
case INT:
case LONG:
case FLOAT:
case DOUBLE: {
ret.isValid = ret.needConvert = isNumeric(readType) || isStringVariant(readType) ||
isDecimal(readType) || isTimestamp(readType);
break;
}
case DECIMAL: {
ret.isValid = ret.needConvert = isNumeric(readType);
break;
}
case STRING:
case CHAR:
case VARCHAR:
case TIMESTAMP:
case TIMESTAMP_INSTANT:
case DATE:
case BINARY: {
// Not support
break;
}
case STRUCT:
case LIST:
case MAP:
case UNION: {
ret.isValid = ret.needConvert = false;
break;
}
default:
break;
}
}
return ret;
}
void SchemaEvolution::buildConversion(const Type* _readType, const Type* fileType) {
if (fileType == nullptr) {
throw SchemaEvolutionError("File does not have " + _readType->toString());
}
auto [valid, convert] = checkConversion(*_readType, *fileType);
if (!valid) {
invalidConversion(_readType, fileType);
}
readTypeMap.emplace(_readType->getColumnId(), convert ? _readType : fileType);
// check whether PPD conversion is safe
buildSafePPDConversionMap(_readType, fileType);
for (uint64_t i = 0; i < _readType->getSubtypeCount(); ++i) {
auto subType = _readType->getSubtype(i);
if (subType) {
// null subType means that this is a sub column of map/list type
// and it does not exist in the file. simply skip it.
buildConversion(subType, fileType->getTypeByColumnId(subType->getColumnId()));
}
}
}
bool SchemaEvolution::needConvert(const Type& fileType) const {
auto _readType = getReadType(fileType);
if (_readType == &fileType) {
return false;
}
// it does not check valid here as verified by buildConversion()
return checkConversion(*_readType, fileType).needConvert;
}
inline bool isPrimitive(const Type* type) {
auto kind = type->getKind();
return kind != STRUCT && kind != MAP && kind != LIST && kind != UNION;
}
void SchemaEvolution::buildSafePPDConversionMap(const Type* _readType, const Type* fileType) {
if (_readType == nullptr || !isPrimitive(_readType) || fileType == nullptr ||
!isPrimitive(fileType)) {
return;
}
bool isSafe = false;
if (_readType == fileType) {
// short cut for same type
isSafe = true;
} else if (_readType->getKind() == DECIMAL && fileType->getKind() == DECIMAL) {
// for decimals alone do equality check to not mess up with precision change
if (fileType->getPrecision() == readType->getPrecision() &&
fileType->getScale() == readType->getScale()) {
isSafe = true;
}
} else {
// only integer and string evolutions are safe
// byte -> short -> int -> long
// string <-> char <-> varchar
// NOTE: Float to double evolution is not safe as floats are stored as
// doubles in ORC's internal index, but when doing predicate evaluation
// for queries like "select * from orc_float where f = 74.72" the constant
// on the filter is converted from string -> double so the precisions will
// be different and the comparison will fail.
// Soon, we should convert all sargs that compare equality between floats
// or doubles to range predicates.
// Similarly string -> char and varchar -> char and vice versa is impossible
// as ORC stores char with padded spaces in its internal index.
switch (fileType->getKind()) {
case BYTE: {
if (readType->getKind() == SHORT || readType->getKind() == INT ||
readType->getKind() == LONG) {
isSafe = true;
}
break;
}
case SHORT: {
if (readType->getKind() == INT || readType->getKind() == LONG) {
isSafe = true;
}
break;
}
case INT: {
if (readType->getKind() == LONG) {
isSafe = true;
}
break;
}
case STRING: {
if (readType->getKind() == VARCHAR) {
isSafe = true;
}
break;
}
case VARCHAR: {
if (readType->getKind() == STRING) {
isSafe = true;
}
break;
}
case BOOLEAN:
case LONG:
case FLOAT:
case DOUBLE:
case BINARY:
case TIMESTAMP:
case LIST:
case MAP:
case STRUCT:
case UNION:
case DECIMAL:
case DATE:
case CHAR:
case TIMESTAMP_INSTANT:
break;
}
}
if (isSafe) {
safePPDConversionMap.insert(fileType->getColumnId());
}
}
bool SchemaEvolution::isSafePPDConversion(uint64_t columnId) const {
return safePPDConversionMap.find(columnId) != safePPDConversionMap.cend();
}
} // namespace orc
|
