使用C ++中的sqlite3命令对表进行排序，添加和剪切

Question

I have a table which looks like this:

id | amount | value
1  |   2    |   3
2  |   4    |   5
3  |   2    |   4

First I would like to sort the table by value:. Then I want to cut of the table when the added amount is greater than a certain max_amount.

Right now I have a routine to do that. How it is supposed to work: It basically looks at each row, compares whether the added_amount is larger than the max_amount. If not it writes the entry in a different, new table if yes it takes the difference (max_amount - added_amount) puts it in the new table and then returns. After that I delete the entries of the old table and transfer the data from the new table into the old one.

RESULT (for max_amount = 5):

id | amount | value
1  |   2    |   3
3  |   2    |   4
2  |   1    |   5

It seems rather complicated and is kind of slow. Therefore, I would love to find an easier way. I was wondering if there was an SQLITE3 command for that or at least a smarter way :)

Hope someone has an idea! :D

~ AMK

EDIT (c++ function would look something like that):

    while(1)
{
    id = id + 1;

    sql = "SELECT * FROM table ORDER BY value ASC LIMIT 1 OFFSET ?";

    sqlite3_stmt *stmt;

    rc = sqlite3_prepare_v2(db, sql, -1, &stmt, NULL);
    if (rc != SQLITE_OK)
    {
        fprintf(stderr, "Can not prepare: %s\n", sqlite3_errmsg(db));
        sqlite3_finalize(stmt);
        exit(0);
    }

    rc = sqlite3_bind_int(stmt, 1, id);
    if (rc != SQLITE_OK) 
    {
        fprintf(stderr, "Can not bind: %s\n", sqlite3_errmsg(db));
        sqlite3_finalize(stmt);
        exit(0);
    }
    rc = sqlite3_step(stmt);
    (...)
    value = sqlite3_column_int(stmt, 0);
    value1 = sqlite3_column_double(stmt, 1);
    value2 = sqlite3_column_double(stmt, 2);
    sqlite3_finalize(stmt);

    sql = "INSERT OR REPLACE INTO table_new (id, amount, value) values (?,?,?)";
    amount_added = amount_added + value1;
    if(amount_added >= amount_max)
    {
        value1 = value1 - (amount_added - amount_max);
                    done = 1;
    }

    rc = sqlite3_prepare_v2(db, sql, -1, &stmt, NULL);
    rc = sqlite3_bind_int(stmt, 0, value);
    rc = sqlite3_bind_int(stmt, 1, value1);
    rc = sqlite3_bind_int(stmt, 2, value2);
    rc = sqlite3_step(stmt);
    sqlite3_finalize(stmt);

    (... put table_new data into table)

            if(done == 1){return = 0;}
}

Answer 1

First of all, take a look at this answer and adapt at your case:

int main() {
  std::string query = "SELECT *  FROM table WHERE amount > 5 ORDER BY value ASC"; 
  system("connectDB.sh " + query.c_str()); 
  /* connectDB.sh should be chmod +x */
}

Your script will be:

#!/bin/bash
sqlite3 test.db ${1}

edit

if you're working on linux platform, you have to create a file, for example:

connectDB.sh

and copy/paste this two rows:

#!/bin/bash
sqlite3 test.db ${1}

first line: you're using bash
second line: execute sqlite3 program on test.db file passing ${1}, the first parameter when you call "connectDB "select* from t"": in this case ${1} = "select* from t"

remember to give to file the execution priviledge:

 chmod +x connectDB.sh

From your main

system("connectDB.sh " + query.c_str()); 

you perform a system call to file connectDB.sh passing the string with your query.

I hope I was clear..

Answer 2

I found out why the routine was rather slow. It is explained in the sqlite-faq https://sqlite.org/faq.html (19). The transaction speed depends on your hard drive:

A transaction normally requires two complete rotations of the disk platter, which on a 7200RPM disk drive limits you to about 60 transactions per second.

Therefore, you can fasten it up using "BEGIN...COMMIT".

Other than that, you can of course combine the two statements into one... With "INSERT OR UPDATE ... SELECT ..." (not really relevant for speed)

I came to the conclusion that you have to use some kind of routine to cut it of and am almost certain, that you can't replace this routine with some kind of sqlite statement.

Thanks for all your help.

~ AMK