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);

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

                string line(linebuf);


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

                        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 << line << endl;
                } else {
                        header << line << 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	#include <iostream>
2	#include <fstream>
3	#include <sstream>
4	#include <string>
5	#include <vector>
6
7
8	#include <boost/algorithm/string.hpp>
9	#include <boost/filesystem.hpp>
10	#include <boost/program_options.hpp>
11
12	namespace po = boost::program_options;
13
14	using namespace std;
15
16	const int BUFSIZE = 1024*1024;
17
18	string remove_comments(string in) {
19	bool isComment = false;
20	ostringstream out;
21	for (int i=0; i<in.length() -1; i++) {
22	if (in.at(i) == '/' && in.at(i+1) == '*')
23	isComment = true;
24
25	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	string get_db_name(string input) {
39	input = remove_comments(input);
40
41	boost::erase_all(input,"`");
42	boost::erase_all(input, ";");
43	boost::trim(input);
44
45	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
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	return 1;
82	}
83
84
85	char linebuf[BUFSIZE];
86
87	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
94	ifstream in( inputfile.c_str() );
95	ofstream out;
96
97	ostringstream header;
98
99	if (!in.is_open() ) {
100	cout << "Could not open " << argv[1] << endl;
101	return 2;
102	}
103
104	time_t start = time(NULL);
105
106	while ( in.good() ) {
107	in.getline(linebuf, BUFSIZE);
108
109	string line(linebuf);
110
111
112	//if (line.substr(0, 15) == "CREATE DATABASE" ) {
113	if ( boost::starts_with(line, "CREATE DATABASE") ) {
114
115	if (out.is_open() ) {
116	out.close();
117	}
118
119	boost::trim(line);
120	string dbname = get_db_name(line);
121	cout << ">" << dbname << endl;
122
123
124	ostringstream oss;
125	oss << dbname << "_dump.sql";
126
127	string filename = oss.str();
128
129	bool did_exist = boost::filesystem::exists(filename);
130
131	out.open( filename.c_str(), ios::app );
132
133
134
135	if (!out.is_open() ) {
136	cout << "could not create outfile " << filename << endl;
137	return 3;
138	}
139
140	if (!did_exist) {
141	out << header.str() << endl; //write preamble in new file
142	}
143
144	}
145
146
147	if (out.is_open() ) {
148	out << line << endl;
149	} else {
150	header << line << endl; //collect preamble for later use
151	}
152
153	}
154
155
156	if(out.is_open())
157	out.close();
158
159	time_t end = time(NULL);
160
161	cout << "Elapsed " << (end-start) << " seconds" << endl;
162
163
164	in.close();
165	}