mlisp/usql/csvreader.cpp

151 lines
3.1 KiB
C++

#include <cerrno>
#include "exception.h"
#include "csvreader.h"
#include "parser.h"
namespace usql {
CsvReader::CsvReader(bool skip_hdr, char field_sep, char quote_ch, char line_sep, char line_sep2) {
skip_header = skip_hdr;
field_separator = field_sep;
quote_character = quote_ch;
line_separator = line_sep;
line_separator2 = line_sep2;
header_skiped = !skip_hdr;
}
int CsvReader::parseCSV(const std::string &filename, std::vector<ColDefNode> &cols_def, Table &table) {
std::vector<ColDefNode> cdefs;
cdefs.reserve(cols_def.size());
for (auto &cd : cols_def) {
cdefs.emplace_back(table.get_column_def(cd.name));
}
int row_cnt = 0;
bool inQuote(false);
std::string field;
std::vector<std::string> line;
line.reserve(32);
errno = 0;
FILE* fp = fopen(filename.c_str(), "r");
if (fp == NULL)
throw Exception("load from csv, cannot read from file(" + std::to_string(errno) + ")");
char* line_str = NULL;
size_t len = 0;
long read_chars;
while ((read_chars = getline(&line_str, &len, fp)) != -1) {
if (skip_header && !header_skiped) {
header_skiped = true;
continue;
}
if (read_chars > 0 && line_str[read_chars - 1] == '\n') {
line_str[read_chars - 1] = '\0';
--read_chars;
}
std::string csvSource{line_str};
std::string::const_iterator aChar = csvSource.begin();
std::string::const_iterator strEnd = csvSource.end();
while (aChar != strEnd) {
if (*aChar == quote_character) {
inQuote = !inQuote;
} else if (*aChar == field_separator) {
if (inQuote) {
field += *aChar;
} else {
line.push_back(field);
field.clear();
}
} else {
field.push_back(*aChar);
}
++aChar;
}
if (!field.empty())
line.push_back(field);
table.create_row_from_vector(cols_def, line);
row_cnt++;
field.clear();
line.clear();
}
fclose(fp);
if (line_str)
free(line_str);
return row_cnt;
}
int CsvReader::parseCSV2(const std::string &csvSource, std::vector<ColDefNode> &cols_def, Table& table) {
int row_cnt = 0;
bool inQuote(false);
bool newLine(false);
std::string field;
std::vector<std::string> line;
line.reserve(32);
std::string::const_iterator aChar = csvSource.begin();
while (aChar != csvSource.end()) {
if (*aChar == quote_character) {
newLine = false;
inQuote = !inQuote;
} else if (*aChar == field_separator) {
newLine = false;
if (inQuote) {
field += *aChar;
} else {
line.push_back(field);
field.clear();
}
} else if (*aChar == line_separator || *aChar == line_separator2) {
if (inQuote) {
field += *aChar;
} else {
if (!newLine) {
line.push_back(field);
if (header_skiped) {
table.create_row_from_vector(cols_def, line);
row_cnt++;
}
header_skiped = true;
field.clear();
line.clear();
newLine = true;
}
}
} else {
newLine = false;
field.push_back(*aChar);
}
aChar++;
}
if (!field.empty()) line.push_back(field);
if (header_skiped) {
table.create_row_from_vector(cols_def, line);
row_cnt++;
header_skiped = true;
}
return row_cnt;
}
} // namespace