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
|
#include "CapnProtoRowInputFormat.h"
#if USE_CAPNP
#include <IO/ReadBuffer.h>
#include <Interpreters/Context.h>
#include <Formats/FormatFactory.h>
#include <Formats/FormatSchemaInfo.h>
#error #include <capnp/serialize.h>
#error #include <capnp/dynamic.h>
#error #include <capnp/common.h>
namespace DB
{
namespace ErrorCodes
{
extern const int INCORRECT_DATA;
}
CapnProtoRowInputFormat::CapnProtoRowInputFormat(ReadBuffer & in_, Block header_, Params params_, const CapnProtoSchemaInfo & info, const FormatSettings & format_settings)
: IRowInputFormat(std::move(header_), in_, std::move(params_))
, parser(std::make_shared<CapnProtoSchemaParser>())
{
// Parse the schema and fetch the root object
schema = parser->getMessageSchema(info.getSchemaInfo());
const auto & header = getPort().getHeader();
serializer = std::make_unique<CapnProtoSerializer>(header.getDataTypes(), header.getNames(), schema, format_settings.capn_proto);
}
std::pair<kj::Array<capnp::word>, size_t> CapnProtoRowInputFormat::readMessagePrefix()
{
uint32_t segment_count;
in->readStrict(reinterpret_cast<char*>(&segment_count), sizeof(uint32_t));
/// Don't allow large amount of segments as it's done in capnproto library:
/// https://github.com/capnproto/capnproto/blob/931074914eda9ca574b5c24d1169c0f7a5156594/c%2B%2B/src/capnp/serialize.c%2B%2B#L181
/// Large amount of segments can indicate that corruption happened.
if (segment_count >= 512)
throw Exception(ErrorCodes::INCORRECT_DATA, "Message has too many segments. Most likely, data was corrupted");
// one for segmentCount and one because segmentCount starts from 0
const auto prefix_size = (2 + segment_count) * sizeof(uint32_t);
const auto words_prefix_size = (segment_count + 1) / 2 + 1;
auto prefix = kj::heapArray<capnp::word>(words_prefix_size);
auto prefix_chars = prefix.asChars();
::memcpy(prefix_chars.begin(), &segment_count, sizeof(uint32_t));
// read size of each segment
for (size_t i = 0; i <= segment_count; ++i)
in->readStrict(prefix_chars.begin() + ((i + 1) * sizeof(uint32_t)), sizeof(uint32_t));
return {std::move(prefix), prefix_size};
}
kj::Array<capnp::word> CapnProtoRowInputFormat::readMessage()
{
auto [prefix, prefix_size] = readMessagePrefix();
auto prefix_chars = prefix.asChars();
// calculate size of message
const auto expected_words = capnp::expectedSizeInWordsFromPrefix(prefix);
const auto expected_bytes = expected_words * sizeof(capnp::word);
const auto data_size = expected_bytes - prefix_size;
auto msg = kj::heapArray<capnp::word>(expected_words);
auto msg_chars = msg.asChars();
// read full message
::memcpy(msg_chars.begin(), prefix_chars.begin(), prefix_size);
in->readStrict(msg_chars.begin() + prefix_size, data_size);
return msg;
}
void CapnProtoRowInputFormat::skipMessage()
{
auto [prefix, prefix_size] = readMessagePrefix();
// calculate size of message
const auto expected_bytes = capnp::expectedSizeInWordsFromPrefix(prefix) * sizeof(capnp::word);
const auto data_size = expected_bytes - prefix_size;
// skip full message
in->ignore(data_size);
}
bool CapnProtoRowInputFormat::readRow(MutableColumns & columns, RowReadExtension &)
{
if (in->eof())
return false;
try
{
auto array = readMessage();
capnp::FlatArrayMessageReader msg(array);
auto root_reader = msg.getRoot<capnp::DynamicStruct>(schema);
serializer->readRow(columns, root_reader);
}
catch (const kj::Exception & e)
{
throw Exception(ErrorCodes::INCORRECT_DATA, "Cannot read row: {}", e.getDescription().cStr());
}
return true;
}
size_t CapnProtoRowInputFormat::countRows(size_t max_block_size)
{
size_t num_rows = 0;
while (!in->eof() && num_rows < max_block_size)
{
skipMessage();
++num_rows;
}
return num_rows;
}
CapnProtoSchemaReader::CapnProtoSchemaReader(const FormatSettings & format_settings_) : format_settings(format_settings_)
{
}
NamesAndTypesList CapnProtoSchemaReader::readSchema()
{
auto schema_info = FormatSchemaInfo(
format_settings.schema.format_schema,
"CapnProto",
true,
format_settings.schema.is_server,
format_settings.schema.format_schema_path);
auto schema_parser = CapnProtoSchemaParser();
auto schema = schema_parser.getMessageSchema(schema_info);
return capnProtoSchemaToCHSchema(schema, format_settings.capn_proto.skip_fields_with_unsupported_types_in_schema_inference);
}
void registerInputFormatCapnProto(FormatFactory & factory)
{
factory.registerInputFormat(
"CapnProto",
[](ReadBuffer & buf, const Block & sample, IRowInputFormat::Params params, const FormatSettings & settings)
{
return std::make_shared<CapnProtoRowInputFormat>(
buf,
sample,
std::move(params),
CapnProtoSchemaInfo(settings, "CapnProto", sample, settings.capn_proto.use_autogenerated_schema),
settings);
});
factory.markFormatSupportsSubsetOfColumns("CapnProto");
factory.registerFileExtension("capnp", "CapnProto");
factory.registerAdditionalInfoForSchemaCacheGetter(
"CapnProto",
[](const FormatSettings & settings)
{
return fmt::format(
"format_schema={}, skip_fields_with_unsupported_types_in_schema_inference={}",
settings.schema.format_schema,
settings.capn_proto.skip_fields_with_unsupported_types_in_schema_inference);
});
}
void registerCapnProtoSchemaReader(FormatFactory & factory)
{
factory.registerExternalSchemaReader("CapnProto", [](const FormatSettings & settings)
{
return std::make_shared<CapnProtoSchemaReader>(settings);
});
}
}
#else
namespace DB
{
class FormatFactory;
void registerInputFormatCapnProto(FormatFactory &) {}
void registerCapnProtoSchemaReader(FormatFactory &) {}
}
#endif // USE_CAPNP
|
