support also attribute metadata (i.e. also different attribute types than mere strings) and empty relations (no records)
--- a/src/YAMLCommand.h Thu Dec 03 22:31:03 2020 +0100
+++ b/src/YAMLCommand.h Sat Dec 05 19:17:21 2020 +0100
@@ -50,6 +50,26 @@
return event.type;
}
+ const relpipe::writer::string_t getTypeName() const {
+ return findTypeName(event.type);
+ }
+
+ /** Just for debugging and error handling */
+ static const relpipe::writer::string_t findTypeName(const yaml_event_type_t eventType) {
+ if (eventType == YAML_NO_EVENT) return L"NO";
+ else if (eventType == YAML_STREAM_START_EVENT) return L"STREAM_START";
+ else if (eventType == YAML_STREAM_END_EVENT) return L"STREAM_END";
+ else if (eventType == YAML_DOCUMENT_START_EVENT) return L"DOCUMENT_START";
+ else if (eventType == YAML_DOCUMENT_END_EVENT) return L"DOCUMENT_END";
+ else if (eventType == YAML_ALIAS_EVENT) return L"ALIAS";
+ else if (eventType == YAML_SCALAR_EVENT) return L"SCALAR";
+ else if (eventType == YAML_SEQUENCE_START_EVENT) return L"SEQUENCE_START";
+ else if (eventType == YAML_SEQUENCE_END_EVENT) return L"SEQUENCE_END";
+ else if (eventType == YAML_MAPPING_START_EVENT) return L"MAPPING_START";
+ else if (eventType == YAML_MAPPING_END_EVENT) return L"MAPPING_END";
+ else return L"UNKNOWN";
+ }
+
const yaml_event_t* getEvent() const {
return &event;
}
@@ -94,31 +114,113 @@
}
};
+ using YAMLEvent_p = std::shared_ptr<YAMLEvent>;
+
YAMLParser parser;
-
- using YAMLEvent_p = std::shared_ptr<YAMLEvent>;
+ std::shared_ptr<relpipe::writer::RelationalWriter> writer;
- enum class State {
- START,
- RELATIONS,
- ATTRIBUTE,
- RECORDS
- };
+ relpipe::writer::string_t relationName;
+ std::vector<relpipe::writer::string_t> record;
+ std::vector<relpipe::writer::AttributeMetadata> attributesMetadata;
- State state;
-
- relpipe::writer::string_t y2s(yaml_char_t* value) {
+ relpipe::writer::string_t y2s(const yaml_char_t* value) {
return value ? convertor.from_bytes((const char*) value) : L"";
}
- relpipe::writer::string_t findValue(std::vector<relpipe::writer::string_t> record, relpipe::writer::string_t key) {
- if (record.size() % 2) relpipe::writer::RelpipeWriterException(L"Invalid count of values in the record vector");
- for (size_t i = 0; i < record.size(); i += 2) {
- if (record[i] == key) return record[i + 1];
+ relpipe::writer::string_t fetchScalarValue(YAMLEvent_p event) {
+ if (event->getType() == YAML_SCALAR_EVENT) return y2s(event->getEvent()->data.scalar.value);
+ else throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: expected SCALAR but got: " + event->getTypeName());
+ }
+
+ void consumeEvent(const yaml_event_type_t expectedEventType, relpipe::writer::string_t expectedScalarValue = L"") {
+ YAMLEvent_p event = YAMLEvent_p(parser.next());
+ if (!event) throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: missing event: " + YAMLEvent::findTypeName(expectedEventType));
+ if (event->getType() != expectedEventType) throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: expected event: " + YAMLEvent::findTypeName(expectedEventType) + L", but got: " + event->getTypeName());
+ if (expectedEventType == YAML_SCALAR_EVENT && expectedScalarValue.size() && expectedScalarValue != fetchScalarValue(event)) throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: expected scalar value: " + expectedScalarValue + L", but got " + fetchScalarValue(event));
+ }
+
+ relpipe::writer::string_t consumeScalarEvent() {
+ YAMLEvent_p event = YAMLEvent_p(parser.next());
+ if (event && event->getType() == YAML_SCALAR_EVENT) return fetchScalarValue(event);
+ else throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: expected SCALAR, but got: " + event->getTypeName());
+ }
+
+ void processRelation() {
+ YAMLEvent_p event = YAMLEvent_p(parser.next());
+ if (event->getType() == YAML_MAPPING_START_EVENT) processRelationWithMetadata();
+ else if (event->getType() == YAML_SEQUENCE_START_EVENT)processRelationWithoutMetadata();
+ else throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: expected MAPPING or SEQUENCE, but got: " + event->getTypeName());
+ }
+
+ void processRelationWithMetadata() {
+ consumeEvent(YAML_SCALAR_EVENT, L"attribute-metadata");
+ consumeEvent(YAML_SEQUENCE_START_EVENT);
+
+ for (YAMLEvent_p event = YAMLEvent_p(parser.next()); event && event->getType() != YAML_SEQUENCE_END_EVENT; event = YAMLEvent_p(parser.next())) {
+ if (event->getType() != YAML_MAPPING_START_EVENT) throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: expected MAPPING (attribute-metadata), but got: " + event->getTypeName());
+ relpipe::writer::string_t name;
+ relpipe::writer::string_t type = L"string";
+ for (YAMLEvent_p event = YAMLEvent_p(parser.next()); event && event->getType() != YAML_MAPPING_END_EVENT; event = YAMLEvent_p(parser.next())) {
+ auto key = fetchScalarValue(event);
+ auto value = consumeScalarEvent();
+ if (key == L"name") name = value;
+ else if (key == L"type") type = value;
+ else; // unsupported metadata, later there might be something useful
+ }
+ attributesMetadata.push_back({name, writer->toTypeId(type)});
}
- return L"";
+
+ writer->startRelation(relationName, attributesMetadata, true);
+
+ YAMLEvent_p event = YAMLEvent_p(parser.next());
+ if (event->getType() == YAML_SCALAR_EVENT && fetchScalarValue(event) == L"record") {
+ consumeEvent(YAML_SEQUENCE_START_EVENT);
+ processRecords();
+ consumeEvent(YAML_MAPPING_END_EVENT);
+ } else if (event->getType() == YAML_MAPPING_END_EVENT) {
+ // empty relation, no records
+ } else {
+ relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: expected 'record' or MAPPING_END, but got: " + event->getTypeName());
+ }
}
+ void processRelationWithoutMetadata() {
+ // First record:
+ consumeEvent(YAML_MAPPING_START_EVENT);
+ record.clear();
+ for (YAMLEvent_p event = YAMLEvent_p(parser.next()); event && event->getType() != YAML_MAPPING_END_EVENT; event = YAMLEvent_p(parser.next())) {
+ auto name = fetchScalarValue(event);
+ auto value = consumeScalarEvent();
+ attributesMetadata.push_back({name, relpipe::writer::TypeId::STRING});
+ record.push_back(value);
+ }
+ writer->startRelation(relationName, attributesMetadata, true);
+ for (auto value : record) writer->writeAttribute(value);
+
+ // Following records:
+ processRecords();
+ }
+
+ void processRecords() {
+ for (YAMLEvent_p event = YAMLEvent_p(parser.next()); event && event->getType() != YAML_SEQUENCE_END_EVENT; event = YAMLEvent_p(parser.next())) {
+ if (event->getType() != YAML_MAPPING_START_EVENT) throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: expected MAPPING (record), but got: " + event->getTypeName());
+ record.clear();
+ record.resize(attributesMetadata.size());
+ for (YAMLEvent_p event = YAMLEvent_p(parser.next()); event && event->getType() != YAML_MAPPING_END_EVENT; event = YAMLEvent_p(parser.next())) {
+ auto name = fetchScalarValue(event);
+ auto value = consumeScalarEvent();
+ for (int i = 0; i < attributesMetadata.size(); i++) {
+ if (name == attributesMetadata[i].attributeName) {
+ record[i] = value;
+ break;
+ }
+ }
+ }
+ for (auto value : record) writer->writeAttribute(value);
+ }
+ }
+
+
public:
YAMLCommand() {
@@ -129,61 +231,20 @@
void process(std::istream& input, std::shared_ptr<relpipe::writer::RelationalWriter> writer) {
parser.setInput(&input);
+ this->writer = writer;
- relpipe::writer::string_t relationName;
- std::vector<relpipe::writer::string_t> record;
- std::vector<relpipe::writer::AttributeMetadata> attributesMetadata;
-
- state = State::START;
+ consumeEvent(YAML_STREAM_START_EVENT);
+ consumeEvent(YAML_DOCUMENT_START_EVENT);
+ consumeEvent(YAML_MAPPING_START_EVENT);
- for (YAMLEvent_p event = YAMLEvent_p(parser.next()); event; event = YAMLEvent_p(parser.next())) {
- if (event->getType() == YAML_NO_EVENT) {
- throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: YAML_NO_EVENT");
- } else if (event->getType() == YAML_STREAM_START_EVENT) {
- } else if (event->getType() == YAML_STREAM_END_EVENT) {
- } else if (event->getType() == YAML_DOCUMENT_START_EVENT) {
- } else if (event->getType() == YAML_DOCUMENT_END_EVENT) {
- } else if (event->getType() == YAML_ALIAS_EVENT) {
- } else if (event->getType() == YAML_SCALAR_EVENT) {
- relpipe::writer::string_t scalarValue = y2s(event->getEvent()->data.scalar.value);
- if (state == State::RELATIONS) relationName = scalarValue;
- else if (state == State::ATTRIBUTE) record.push_back(scalarValue);
- else throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: unexpected YAML_SCALAR_EVENT");
- } else if (event->getType() == YAML_SEQUENCE_START_EVENT) {
- if (state == State::RELATIONS) state = State::RECORDS;
- else throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: unexpected YAML_SEQUENCE_START_EVENT");
- } else if (event->getType() == YAML_SEQUENCE_END_EVENT) {
- if (state == State::RECORDS) {
- state = State::RELATIONS;
- relationName.clear();
- attributesMetadata.clear();
- } else {
- throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: unexpected YAML_SEQUENCE_END_EVENT");
- }
- } else if (event->getType() == YAML_MAPPING_START_EVENT) {
- if (state == State::START) state = State::RELATIONS;
- else if (state == State::RECORDS) state = State::ATTRIBUTE;
- else if (state == State::RELATIONS) throw relpipe::writer::RelpipeWriterException(L"Not yet implemented"); // TODO: there might be also a map in the relation value (not only a sequence) that would contain metadata (i.e. data types + support for relation containing no records)
- else throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: unexpected YAML_MAPPING_START_EVENT");
- } else if (event->getType() == YAML_MAPPING_END_EVENT) {
- if (state == State::ATTRIBUTE) {
- if (attributesMetadata.size() == 0) {
- if (record.size() % 2) relpipe::writer::RelpipeWriterException(L"Invalid count of values in the record vector");
- for (size_t i = 0; i < record.size(); i += 2) attributesMetadata.push_back(relpipe::writer::AttributeMetadata{record[i], relpipe::writer::TypeId::STRING});
- writer->startRelation(relationName, attributesMetadata, true);
- }
- for (auto m : attributesMetadata) writer->writeAttribute(findValue(record, m.attributeName));
- record.clear();
- state = State::RECORDS;
- } else if (state == State::RELATIONS) {
- break; // map of the relations ends (root)
- } else {
- throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: unexpected YAML_MAPPING_END_EVENT");
- }
- } else {
- throw relpipe::writer::RelpipeWriterException(L"Invalid YAML structure: unknown event: " + std::to_wstring(event->getType()));
- }
+ for (YAMLEvent_p event = YAMLEvent_p(parser.next()); event && event->getType() != YAML_MAPPING_END_EVENT; event = YAMLEvent_p(parser.next())) {
+ relationName = fetchScalarValue(event);
+ attributesMetadata.clear();
+ processRelation();
}
+
+ consumeEvent(YAML_DOCUMENT_END_EVENT);
+ consumeEvent(YAML_STREAM_END_EVENT);
}
};