misc/mysql_splitter/splitter.cpp

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

#include <cstdio>

#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 (unsigned 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;
        }


        FILE* in = fopen64( inputfile.c_str(), "r" );
        FILE* out = NULL;
        //ifstream in( inputfile.c_str() );
        //ofstream out;

        ostringstream header;

        if ( in == NULL ) {
                cout << "Could not open " << argv[1] << endl;
                perror("");
                return 2;
        }

        time_t start = time(NULL);

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

        while ( feof(in) == 0 && ferror(in) == 0 ) {
                fgets(linebuf, BUFSIZE, in);


                //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 != NULL ) {
                                fclose(out);
                        }

                        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 );
                        out  = fopen64(filename.c_str(), "a");


                        if ( out == NULL ) {
                                cout << "could not create outfile " << filename << endl;
                                return 3;
                        }

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

                }


                if (out != NULL ) {
                        fputs(linebuf, out);
                        //fputs("\n", out);
                        //out << linebuf << endl;
                } else {
                        header << linebuf; //collect preamble for later use
                }

        }


        if (out != NULL)
                fclose(out); 

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

    time_t end = time(NULL);


        time_t elapsed = end-start;
        cout << setfill('0') << "Elapsed time " << elapsed/60 << ":" << setw(2) << elapsed%60 << endl;

        fclose(in);
}
1	#include <iostream>
2	#include <iomanip>
3	#include <sstream>
4	#include <string>
5	#include <vector>
6
7	#include <cstdio>
8
9	#include <boost/algorithm/string.hpp>
10	#include <boost/filesystem.hpp>
11	#include <boost/program_options.hpp>
12
13	namespace po = boost::program_options;
14
15	using namespace std;
16
17	const int BUFSIZE = 1024*1024;
18
19	string remove_comments(string in) {
20	bool isComment = false;
21	ostringstream out;
22	for (unsigned int i=0; i<in.length() -1; i++) {
23	if (in.at(i) == '/' && in.at(i+1) == '*')
24	isComment = true;
25
26	if (in.at(i) == '*' && in.at(i+1) == '/') {
27	i++;
28	isComment = false;
29	continue;
30	}
31
32	if (isComment == false)
33	out << in.at(i) ;
34	}
35
36	return out.str();
37	}
38
39	string get_db_name(string input) {
40	input = remove_comments(input);
41
42	boost::erase_all(input,"`");
43	boost::erase_all(input, ";");
44	boost::trim(input);
45
46	vector<string> words;
47	words = boost::split(words, input, boost::is_any_of(" ") );
48
49	string last = words.back();
50	boost::trim(last);
51
52	return last;
53	}
54
55
56	int main(int argc, char** argv) {
57
58	po::options_description desc("Mysql dump file splitter:\nAllowed options");
59	desc.add_options()
60	("help", "produce help message")
61	("data", "ignore ceate tablestatements")
62	("input-file", po::value< string >(), "input file")
63	;
64
65	po::positional_options_description p;
66	p.add("input-file", -1);
67
68
69	po::variables_map vm;
70	po::store( po::command_line_parser(argc, argv).options(desc).positional(p).run(), vm);
71	po::notify(vm);
72
73	if (vm.count("help")) {
74	cout << desc << "\n";
75	return 1;
76	}
77
78
79
80	if ( vm.count("input-file") == 0) {
81	cout << desc << endl;
82	return 1;
83	}
84
85
86	char linebuf[BUFSIZE];
87
88	string inputfile = vm["input-file"].as< string >();
89	if (! boost::filesystem::exists(inputfile) ) {
90	cout << "No file named " << inputfile << endl;
91	return 1;
92	}
93
94
95	FILE* in = fopen64( inputfile.c_str(), "r" );
96	FILE* out = NULL;
97	//ifstream in( inputfile.c_str() );
98	//ofstream out;
99
100	ostringstream header;
101
102	if ( in == NULL ) {
103	cout << "Could not open " << argv[1] << endl;
104	perror("");
105	return 2;
106	}
107
108	time_t start = time(NULL);
109
110	const char* SEARCH = "CREATE DATABASE";
111	const int SEARCHLEN = strlen(SEARCH);
112
113	while ( feof(in) == 0 && ferror(in) == 0 ) {
114	fgets(linebuf, BUFSIZE, in);
115
116
117	//if (line.substr(0, 15) == "CREATE DATABASE" ) {
118	//if ( boost::starts_with(line, "CREATE DATABASE") ) {
119	if (strncmp(linebuf,SEARCH, SEARCHLEN) == 0) {
120
121	string line(linebuf);
122
123	if (out != NULL ) {
124	fclose(out);
125	}
126
127	boost::trim(line);
128	string dbname = get_db_name(line);
129	cout << ">" << dbname << endl;
130
131
132	ostringstream oss;
133	oss << dbname << "_dump.sql";
134
135	string filename = oss.str();
136
137	bool did_exist = boost::filesystem::exists(filename);
138
139	//out.open( filename.c_str(), ios::app );
140	out = fopen64(filename.c_str(), "a");
141
142
143
144	if ( out == NULL ) {
145	cout << "could not create outfile " << filename << endl;
146	return 3;
147	}
148
149	if (!did_exist) {
150	fputs(header.str().c_str(), out);
151	//out << header.str() << endl; //write preamble in new file
152	}
153
154	}
155
156
157	if (out != NULL ) {
158	fputs(linebuf, out);
159	//fputs("\n", out);
160	//out << linebuf << endl;
161	} else {
162	header << linebuf; //collect preamble for later use
163	}
164
165	}
166
167
168	if (out != NULL)
169	fclose(out);
170
171	// if (out.is_open())
172	// out.close();
173
174	time_t end = time(NULL);
175
176
177	time_t elapsed = end-start;
178	cout << setfill('0') << "Elapsed time " << elapsed/60 << ":" << setw(2) << elapsed%60 << endl;
179
180	fclose(in);
181	}