first working version v_0
authorFrantišek Kučera <franta-hg@frantovo.cz>
Wed, 09 Jan 2019 23:04:51 +0100
branchv_0
changeset 1 5eb4d149c6e2
parent 0 eca0b23802e8
child 2 e83895da3e8f
first working version
src/relpipe-in-csv.cpp
--- a/src/relpipe-in-csv.cpp	Wed Jan 09 17:23:02 2019 +0100
+++ b/src/relpipe-in-csv.cpp	Wed Jan 09 23:04:51 2019 +0100
@@ -16,7 +16,7 @@
  * along with this program. If not, see <http://www.gnu.org/licenses/>.
  */
 #include <cstdlib>
-#include <fstream>
+#include <vector>
 #include <memory>
 #include <regex>
 #include <algorithm>
@@ -24,6 +24,7 @@
 
 #include <relpipe/writer/RelationalWriter.h>
 #include <relpipe/writer/RelpipeWriterException.h>
+#include <relpipe/writer/AttributeMetadata.h>
 #include <relpipe/writer/Factory.h>
 #include <relpipe/writer/TypeId.h>
 
@@ -33,13 +34,79 @@
 using namespace relpipe::cli;
 using namespace relpipe::writer;
 
-/**
- * see https://hg.frantovo.cz/sql-api/file/tip/prototyp/prototyp.sql#l49
- */
-void processDataStream(ostream &output, istream* input) {
+bool readValue(istream& input, stringstream& currentValue, bool& lastInRecord) {
+	lastInRecord = false;
+	char ch;
+	input.get(ch);
+	if (ch == '"') {
+		while (input.get(ch)) {
+			if (ch == '"') {
+				input.get(ch);
+				if (ch == '"') {
+					currentValue << ch;
+				} else {
+					if (ch == '\r') input.get(ch);
+					if (ch == '\n') lastInRecord = true;
+					else if (ch != ',') throw RelpipeWriterException(L"Unexpected character (should be ā€ž\\nā€œ or ā€ž,ā€œ)");
+					return true;
+				}
+			} else {
+				currentValue << ch;
+			}
+		}
+	} else if (ch == ',') {
+		return true;
+	} else if (ch == '\n') {
+		lastInRecord = true;
+		return true;
+	} else if (ch == '\r') {
+		currentValue << ch;
+		if (ch == '\n') {
+			lastInRecord = true;
+			return true;
+		} else {
+			throw RelpipeWriterException(L"Crazy carriage stuck during journey");
+		}
+	} else {
+		for (currentValue << ch; input.get(ch);) {
+			switch (ch) {
+				case ',': return true;
+				case '\r': break;
+				case '\n':
+					lastInRecord = true;
+					return true;
+				default: currentValue << ch;
+			}
+		}
+	}
+	return false;
+}
 
-	for (char ch; input->get(ch);) {
-		output << "ch: " << ch << endl;
+void processDataStream(ostream &output, istream& input) {
+	wstring_convert < codecvt_utf8<wchar_t>> convertor; // UTF-8 is required for CSV
+	std::shared_ptr<RelationalWriter> writer(Factory::create(output));
+	vector<AttributeMetadata> metadata;
+	bool headerDone = false;
+	bool lastInRecord = false;
+	stringstream currentValue;
+
+
+	while (readValue(input, currentValue, lastInRecord) && input.good()) {
+		if (headerDone) {
+			writer->writeAttribute(convertor.from_bytes(currentValue.str()));
+		} else {
+			AttributeMetadata am;
+			am.attributeName = convertor.from_bytes(currentValue.str());
+			am.typeId = TypeId::STRING; // TODO: support also other types passed as CLI parameters
+			metadata.push_back(am);
+			if (lastInRecord) {
+				headerDone = true;
+				writer->startRelation(L"csv", metadata, true); // TODO: support also custom relation name passed as CLI parameter
+			}
+		}
+
+		currentValue.str("");
+		currentValue.clear();
 	}
 }
 
@@ -51,7 +118,7 @@
 	int resultCode = CLI::EXIT_CODE_UNEXPECTED_ERROR;
 
 	try {
-		processDataStream(cout, &cin);
+		processDataStream(cout, cin);
 		resultCode = CLI::EXIT_CODE_SUCCESS;
 	} catch (RelpipeWriterException e) {
 		fwprintf(stderr, L"Caught Writer exception: %ls\n", e.getMessge().c_str());