misc/mysql_splitter/splitter.cpp

#include <iostream>
#include <fstream>
#include <sstream>
#include <string>
#include <vector>


#include <boost/algorithm/string.hpp>
#include <boost/filesystem.hpp>
#include <boost/program_options.hpp>

namespace po = boost::program_options;

using namespace std;

const int BUFSIZE = 1024*1024;

string remove_comments(string in) {
        bool isComment = false;
        ostringstream out;
        for (int i=0; i<in.length() -1; i++) {
                if (in.at(i) == '/' && in.at(i+1) == '*')
                        isComment = true;

                if (in.at(i) == '*' && in.at(i+1) == '/') {
                        i++;
                        isComment = false;
                        continue;
                }

                if (isComment == false)
                        out << in.at(i) ;
        }

        return out.str();
}

string get_db_name(string input) {
        input = remove_comments(input);

        boost::erase_all(input,"`");
        boost::erase_all(input, ";");
        boost::trim(input);

        vector<string> words;
        words = boost::split(words, input, boost::is_any_of(" ") );

        string last = words.back();
        boost::trim(last);

        return last;
}


int main(int argc, char** argv) {

        po::options_description desc("Mysql dump file splitter:\nAllowed options");
        desc.add_options()
        ("help", "produce help message")
        ("data", "ignore ceate tablestatements")
                ("input-file", po::value< string >(), "input file")
        ;

        po::positional_options_description p;
        p.add("input-file", -1);


        po::variables_map vm;
        po::store( po::command_line_parser(argc, argv).options(desc).positional(p).run(), vm);
        po::notify(vm);

        if (vm.count("help")) {
        cout << desc << "\n";
        return 1;
        }


        if ( vm.count("input-file") == 0) {
                cout << desc << endl;
                return 1;
        }


        char linebuf[BUFSIZE];

        string inputfile = vm["input-file"].as< string >();
        if (! boost::filesystem::exists(inputfile) ) {
                cout << "No file named " << inputfile << endl;
                return 1;
        }


        ifstream in( inputfile.c_str() );
        ofstream out;

        ostringstream header;

        if (!in.is_open() ) {
                cout << "Could not open " << argv[1] << endl;
                return 2;
        }

        time_t start = time(NULL);

        const char* SEARCH = "CREATE DATABASE";
        const int SEARCHLEN = strlen(SEARCH);

        while ( in.good() ) {
                in.getline(linebuf, BUFSIZE);


                //if (line.substr(0, 15) == "CREATE DATABASE" ) {
                //if ( boost::starts_with(line, "CREATE DATABASE") ) {
                if (strncmp(linebuf,SEARCH, SEARCHLEN) == 0) {

            string line(linebuf);
                        if (out.is_open() ) {
                                out.close();
                        }

                        boost::trim(line);
                        string dbname = get_db_name(line);
                        cout << ">" << dbname << endl;


                        ostringstream oss;
                        oss << dbname << "_dump.sql";

            string filename = oss.str();

                        bool did_exist = boost::filesystem::exists(filename);

                        out.open( filename.c_str(), ios::app );


                        if (!out.is_open() ) {
                                cout << "could not create outfile " << filename << endl;
                                return 3;
                        }

                        if (!did_exist) {
                                out << header.str() << endl; //write preamble in new file
                        }

                }


                if (out.is_open() ) {
                        out << linebuf << endl;
                } else {
                        header << linebuf << endl; //collect preamble for later use
                }

        }


        if(out.is_open())
                out.close();

    time_t end = time(NULL);

    cout << "Elapsed " << (end-start) << " seconds" << endl;


        in.close();
}
1	torben	1231	#include <iostream>
2			#include <fstream>
3			#include <sstream>
4			#include <string>
5			#include <vector>
6
7	torben	1282
8	torben	1231	#include <boost/algorithm/string.hpp>
9			#include <boost/filesystem.hpp>
10	torben	1282	#include <boost/program_options.hpp>
11	torben	1231
12	torben	1282	namespace po = boost::program_options;
13
14	torben	1231	using namespace std;
15
16			const int BUFSIZE = 1024*1024;
17
18	torben	1283	string remove_comments(string in) {
19			bool isComment = false;
20			ostringstream out;
21			for (int i=0; i<in.length() -1; i++) {
22	torben	1458	if (in.at(i) == '/' && in.at(i+1) == '*')
23	torben	1283	isComment = true;
24	torben	1458
25	torben	1283	if (in.at(i) == '*' && in.at(i+1) == '/') {
26			i++;
27			isComment = false;
28			continue;
29			}
30
31			if (isComment == false)
32			out << in.at(i) ;
33			}
34
35			return out.str();
36			}
37
38	torben	1284	string get_db_name(string input) {
39	torben	1283	input = remove_comments(input);
40
41			boost::erase_all(input,"`");
42	torben	1284	boost::erase_all(input, ";");
43	torben	1283	boost::trim(input);
44	torben	1284
45	torben	1231	vector<string> words;
46			words = boost::split(words, input, boost::is_any_of(" ") );
47
48			string last = words.back();
49			boost::trim(last);
50
51			return last;
52			}
53
54
55			int main(int argc, char** argv) {
56	torben	1282
57			po::options_description desc("Mysql dump file splitter:\nAllowed options");
58			desc.add_options()
59			("help", "produce help message")
60			("data", "ignore ceate tablestatements")
61			("input-file", po::value< string >(), "input file")
62			;
63
64			po::positional_options_description p;
65			p.add("input-file", -1);
66
67
68			po::variables_map vm;
69			po::store( po::command_line_parser(argc, argv).options(desc).positional(p).run(), vm);
70			po::notify(vm);
71
72			if (vm.count("help")) {
73			cout << desc << "\n";
74			return 1;
75			}
76
77
78
79			if ( vm.count("input-file") == 0) {
80			cout << desc << endl;
81	torben	1231	return 1;
82			}
83
84	torben	1282
85	torben	1231	char linebuf[BUFSIZE];
86
87	torben	1282	string inputfile = vm["input-file"].as< string >();
88			if (! boost::filesystem::exists(inputfile) ) {
89			cout << "No file named " << inputfile << endl;
90			return 1;
91			}
92
93	torben	1458
94	torben	1282	ifstream in( inputfile.c_str() );
95	torben	1231	ofstream out;
96
97			ostringstream header;
98
99			if (!in.is_open() ) {
100			cout << "Could not open " << argv[1] << endl;
101			return 2;
102	torben	1458	}
103
104	torben	1459	time_t start = time(NULL);
105
106			const char* SEARCH = "CREATE DATABASE";
107			const int SEARCHLEN = strlen(SEARCH);
108	torben	1231
109			while ( in.good() ) {
110			in.getline(linebuf, BUFSIZE);
111
112
113			//if (line.substr(0, 15) == "CREATE DATABASE" ) {
114	torben	1459	//if ( boost::starts_with(line, "CREATE DATABASE") ) {
115			if (strncmp(linebuf,SEARCH, SEARCHLEN) == 0) {
116	torben	1231
117	torben	1459	string line(linebuf);
118	torben	1231	if (out.is_open() ) {
119			out.close();
120			}
121
122			boost::trim(line);
123	torben	1284	string dbname = get_db_name(line);
124	torben	1231	cout << ">" << dbname << endl;
125
126
127			ostringstream oss;
128			oss << dbname << "_dump.sql";
129
130			string filename = oss.str();
131
132			bool did_exist = boost::filesystem::exists(filename);
133
134			out.open( filename.c_str(), ios::app );
135
136
137
138			if (!out.is_open() ) {
139			cout << "could not create outfile " << filename << endl;
140			return 3;
141			}
142
143			if (!did_exist) {
144			out << header.str() << endl; //write preamble in new file
145			}
146
147			}
148
149
150			if (out.is_open() ) {
151	torben	1459	out << linebuf << endl;
152	torben	1231	} else {
153	torben	1459	header << linebuf << endl; //collect preamble for later use
154	torben	1231	}
155
156			}
157
158
159			if(out.is_open())
160	torben	1458	out.close();
161
162			time_t end = time(NULL);
163
164			cout << "Elapsed " << (end-start) << " seconds" << endl;
165	torben	1231
166
167			in.close();
168			}